Energy-Security Tradeoff in a Secure Cache Architecture Against Buffer Overflow Attacks
- 格式:pdf
- 大小:426.36 KB
- 文档页数:9
电力市场常用词汇(2008-05-29 21:37:49)英文缩写英文全称中文解释ATC Available Transmission Capability 可用输电容量ALM Active Load Management 有功负荷管理Active Power Market 有功市场AS Ancillary service 辅助服务Ancillary service charges 辅助服务费ASM Ancillary Service Market 辅助服务市场AS Ancillary Service Provider 辅助服务供应商Annual peak load curve 年最大负荷曲线Arbitrage 套利ACE Area control error 区域控制误差Area price 分区电价AR Area Regulation 区域管制Auction 拍卖、竞价Auction market 拍卖市场Auction Price 拍卖价格AMP Automated Mitigation Procedures 自动市场势力消除程序AGC Automatic generation control 自动发电控制ADP Availability Declaration Period 可用容量申报期Available capacity 可用发电容量ATC Available transfer capability 可用传输容量B TO B 企业对企业B TOC 企业对客户Bain index 贝恩指数BSC Balancing and settlement code 平衡与结算规则BME Balancing Market Evaluation 平衡市场计算Bid 竞价BS Bid Sufficiecy 申报充足率Bid—based 基于竞价的Bidder’s duration of validity 报价员有效期Bidding Price 竞标价格BPS Bidding Processing System 报价处理系统Bidding unit 报价机组Bids 报价Bilateral contract 双边合同Bilateral market 双边交易市场Bilateral Transaction 双边交易(即大用户直购电)Billing and Accounting 结算和结帐biomass 生物质能Block contracts 分段合同Boundary Flow 界面潮流CBM Capacity Benefit Margin 容量效益裕度Capacity factor 容量系数/容量因子Capacity payment 容量电价Capacity requirement market 容量需求市场Capacity Reservation Tariff 容量预定价格Cap-and-trade 限额交易Clearing 清算Clearing energy 出清电量Cogeneration 热电联产Competitive Electricity Market 竞争性电力市场Congestion 阻塞Congestion Cost 阻塞成本Congestion management 阻塞管理Congestion surplus 阻塞节余Congestion uplift 阻塞上抬费用CUSC Connection and use of system code 并网与网络使用规则Connection charge 并网费Consumer Surplus 消费者剩余CFD Contract for Difference 差价合同CMS Contract Management System 合同管理系统Contract price 合同电量CfDs Contracts for differences 差价和约CMS Contrat Management Subsystem 合同管理子系统Cooperative game 合作博弈Cost of service regulation 服务成本管制Cost of unit start-up 机组启动费用CSP Curtailment Service Provider 缩减负荷提供者Customer baseline load 用户基线负荷day-ahead 日前Decentralized trading model 分散交易模型Default provider 默认服务提供商Demand Curve 需求曲线DTC Demand for Transmission Capacity 输电容量需求DSM Demand-side management 需求侧管理deregulation 市场化dispatch 调度DPF Dispatcher Power Flow 调度员潮流DTS Dispatcher Training Simulator 调度员培训模拟Distribution 配电D Distribution Service Provide 配电商Economic Dispatch 经济调度Electric load management 电力负荷管理EDC Electrical Distribution Company 电力配电公司EFA Electricity Forward Agreement 电能远期合同EMOS Electricity Market Operation System 电力市场运营系统NETA electricity trading arrangement 电力交易协议Eligible customer 合格用户Spot market 现货市场Futures Market 期货交易市场EMC 节能服务公司EMS Energy Management System 能量管理系统EMOS Electricity market operation system 电力市场运营系统RES Renewable sources of energy 可再生能源Energy Spot Market 电能现货市场ERP Enterprise Resource Planning 企业资源计划Equilibrium Price 均衡价格ESCO 能源服务公司ETC Existing Transmission Commit-ments 现存输送协议Ex-post price 事后结算价格Federal Energy Regulatory Commission 联邦能源管理委员会feed-in systems 馈入系统Final closing price 最终收盘价格FTR Financial Transmission Right 金融输电权FTR Firm transmission right 固定输电权FD Fixed demand 固定需求FGR Flow-based Transmission Right 基于潮流的输电权Flowgate 关口输电权(潮流关口)FGR Flowgate Right 关口金融输电权Forced outage 强迫停机Forward Market 远期合同市场Fossil Fuel Levy 矿物燃料税Frequency Control Ancillary Service 频率控制辅助服务Future Market 期货市场Generation 发电量GBS generation bidding system 发电报价系统Generation market 发电市场G Generator 发电商Green certificate market “绿证”市场GC Grid code 输电网技术规范GC Grid Company 电网拥有者GMC Grid Management Charge 电网管理费Hedging 套利保值High and low matching method 高低匹配法hour-ahead 时前Hourly contracts 小时合同Illegal speculation 非法投机行为IMO Independent Market Operator 独立市场运营机构IPP Independent Power Producer 独立发电商ISO Independent System Operator 独立系统操作员Inflexible units 不可调度机组ICAP Installed capacity 装机容量installed generation capacity 装机容量Installed Reserve Margin 装机备用余量integrate resource planning 综合资源规划Integrated trading model 一体化交易模型Intelligent price caps 智能价格上限Interchange price 可中断电价Inter-change Schedule 交易计划Interruptible service 可中断的服务Load characteristics 负荷特性LF Load Forecasting 负荷预测LFC Load Frequency Control 负荷频率控制Load profiling 负荷拟合曲线LSE Load Serving Entity 负荷服务企业LBMP Marginal Price 节点边际电价Location Price 区域价格LMP Locational marginal price 分区边际电价LOLP Loss of Load Probility 电力不足概率losses 网损LVAC Low Voltage Access charge 低压连网费MCP Marginal Clearing Price 市场出清价格Marginal cost pricing 边际成本定价Marginal costs 边际成本Marginal Price 边际电价market Architecture 市场设计MCP Market Clearing Price 市场出清价格Market entry certification system 市场准入制度Market Equilibrium 市场均衡Market failure 市场失灵MMU Market monitoring unit 市场监察机构Market participant 市场参与者/市场成员/市场主体Market power 市场力MSO Market Service Organization 市场管理机构Merit order 优先顺序Merit order price 最优顺序价格Metering Data Acquisition System 电能量采集系统MAAC Mid-Atlantic Area Council 大西洋中部地区委员会Minimum daily load 日最小负荷Model of Transmission Right 输电权模型Monopoly 垄断MSS Multi-Settlement System 多结算系统Multi-part bidding 多部投标MW Daily MW 日报NE Nash equilibrium 纳什均衡network congestion 网络阻塞NETA New Electricity Trading Arrangements 新电力交易规则Nodal price 节点电价Non-bidding unit 非竞价机组NFFO Non-Fossil Fuel Obligation 非矿物燃料契约NERC North American Electric Reliability Council 北美电力可靠性委员会OI Office of Interconnection 办公互联网oligopolistic electricity market 卖方垄断电力市场Open access 开放接入OASIS Open Access Same-Time Information System 输电网实时信息发布系统Operating Agreement 运营协议Operation and Maintain 运行维护OPF Optimal Power Flow 最优潮流ORP Optimal Reactive Power 无功优化Optimization 最优化Option 期权合同Outage Scheduler 检修计划Partial Equilibrium 局部均衡Pattern of Transmission Right 输电权模式PAB Pay-as-bid Settlement 按报价结算Payment 支付PTR physical transmission right 物理输电权Point—to—Point Transmission Right 点对点式输电权Pool 电力库PPP Pool Purchase Price 市场购电价Pool Sale Price 市场售电价PSP Pool Sell Price 市场售电价power Broker 电力经纪人Power Exchange 短期现货交易power market 电力市场PM Power Marketer 发电经纪商Power Pool 电力交易所PPA Power Purchase Agreement 电力购销协议Power wheeling 电力转运Pre-dispatch 预调度Price cap 价格上限Price control 价格控制Price regulation 价格监管Price Sensitive Demand 价格敏感性需求Procurement auction 采购竞价Producer Surplus 生产者剩余Projected Assessment of System Adequacy 市场充裕性评估POLR Provider of last resort 收容性供电商RORC Rate of Response Compliance 负荷响应率Reactive Power Market 无功市场real time market 实时市场real-time balancing market 实时平衡市场Real-time prices 实时电价Reference bid 参考投标/报价RTO Regional Transmission organization 地区输电组织RMCP Regulation Market Clearing Price 调频市场出清价格RAA Reliability Assurance Agreement 可靠性协议RMR Reliability Must Run 由于可靠性原因必须运行机组Renewable geothermal 地热能RPS Renewable portfolio standard 可再生能源份额制标准RSI Residual Supply Index 供给剩余系统Retail access 零售准入Retail market 电力零售市场R Retail Service Provider 零售商SC Schedule Coordinators 计划协调员SCD Security Constrained Dispatch 安全约束经济调度SCUC Security-Constrained Unit Commitment 带安全约束机组组合settlement 结算SBS settlement and billing system 结算系统Settlement Day 结算日Shadow price 影子价格SFT Simultaneous Feasibility Test 可行性测试single buyer 单一买方solar photovoltaics 光电子能solar thermal 太阳能SM Spot Market 现货市场Spot price 现货价格Standard Market 标准电力市场State Estimation 状态估计Stranded costs 搁浅成本Strategic bid 报价策略SCADA Supervisory control and data acquisition 数据采集和监视系统Supply Curve 供给曲线SMP System marginal price 系统边际价格SMV system marginal value 系统边际价值System operator 系统调度机构System price 系统电价Tariff 目录电价VLL the Value of Loss of Load 失负荷电价Top-down pricing 自上而下的定价方法TTC Total transfer capability 最大输电能力Trading arrangement 交易规则Transmission and Ancillary Services Market 输电与辅助服务市场Transmission Congestion 输电阻塞TCCs Transmission Congestion Contracts 输电阻塞合同TCR Transmission Congestion Right 输电阻塞权TLR Transmission Loading Relief 输电负荷切除Transmission network or transmission system 输电网络或输电系统TO Transmission owner 输电业者TRM Transmission Reliability Margin 输电可靠性裕度TMS 交易管理系统TRR Transmission Revenue Requirement 输电收入要求Transmission Right 输电权Unconstrained Schedule 无约束计划UC unit commitment 机组组合UDS Unit Dispatch System 机组调度系统Uplift 上浮电价VLL Value of Lost Load 负荷停电损失Virtual bidding 虚拟投标/报价WSA Weather Sensitive Adjustment 天气—敏感性调整WSCC Western Systems Coordinating Council 西部电力协调委员会Wheeling trading model 过网模式Wholesale market 电力批发市场Zonal pricing 区域定价actual load curve 实际负荷曲线actual active power output curve 实际有功出力曲线actual reactive power output curve 实际无功出力曲线accounting cost 会计成本bid curve 报价曲线biding energy 竞价空间bi-directional contract for difference 双向差价合同bilateral contact for difference 双边差价合同black start service 黑启动服务block biding 分段竞价buying long 买空compesnation regulation 补偿调节competition energy 竞争电量constrained trading schedule 有约束交易计划contract-path method 合同路径法daily load factor 日负荷率day-ahead trading 日前交易DATS day-ahead trade subsystem 日前交易子系统dispatch interval 调度时段dispatch price 调度价格distribution price 配电电价DMIS dispatching management information system 调度管理信息系统efficiency market 有效市场electricity fee 电费electricity market model 电力市场模式electricity matket regulation 电力市场监管elementary ancillary services 基本辅助服务embedded cost pricing 会计成本定价emergency reserve 事故备用energy price 电量电价exclusive service tariff 专项服务价格feasible capacity 可调出力feasible hours 可调小时financial power trading 电力金融交易fixed cost 固定成本floody season-dry season price 丰枯电价frequency curve 频率曲线future contract 期货合同future trading 期货交易generation competition 发电竞争模式generation re-scheduling 发电再计划generation right transfer trading 发电权转让交易generation unit availability 发电机组可用率generation rated capacity 发电机额定容量high-reliability price 高可靠性电价imperfect competitive market 不完全竞争市场inter-connection tariff 联网价IRR internal rate of return 内部收益率lincense system for electric power business 电力业务许可证制度load coincidence factor 负荷同时率load factor 负荷因数load fluctuation 负荷波动long-run marginal cost pricing 长期边际成本定价loss allocation 网损分摊loss conversion 网损折算loss factor 网损系数LTS long-term trade subsystem 长期交易子系统marginal revenue 边际收益marginal utility 边际效用market intervention 市场干预market mechanism 市场机制MAS market analysis subsystem 市场分析子系统multi-block bidding 多段报价multilateral trading 多边交易must-run unit 强制运行机组nonlinear pricing 非线性定价off-grid energy 下网电量on-grid energy 上网电量on-grid price 上网电价one-part price 单一制电价opportunity cost 机会成本peak-valley price 峰谷电价perfect competitive market 完全竞争市场physical power trading 电力实物交易postage stamp method 邮票法power flow tracing method 潮流跟踪法power regulatory agency 电力监管机构power retailer 电力零售商principle of marginality 边际原则reserve capacity 备用容量reserve service 备用服务retail competition 零售竞争模式retail price 销售电价RTS real-time trade subsystem 实时交易子系统scheduled outage 计划停运seasonal price 季节电价sequential bidding 分次竞价single-block bidding 单段报价single-buyer 单一购买者模式time sharing bidding 分时竞价TMR tele-meter reading system 电能量计量系统two-part price 两部制电价uni-directional contract for difference 单项差价合同valid grid assets 电网有效资产variable cost 变动成本verticallty integrated monopoly 垂直垄断模式whole energy competition 批发竞争模式Active regulation 主动监管Annual contract on-grid energy 年度合同上网电量ACRS Accelerated Cost Recovery System 加速成本回收体系Balancing account 平衡帐户Budget line 预算线Capital & interest price 还本付息电价Ceiling and floor of market clearing price 市场出清价格的上下限Deregulation 放松管制Forced outrage 强迫停运Frequency regulation 一次调频Grid access tariff 接入价Grid security assessment 网络安全校核Grid security constraint 网络安全约束Hour-ahead trading 时前交易Indifference curve 无差异曲线Information asymmetry 信息不对称Initial margin 初始保证金Interruptible forward contract 可中断远期合同Interruptible service 可中断服务Investment regulation 投资监管Information publishing subsystem 信息发布子系统Monopoly market 垄断市场Natural monopoly 自然垄断Nodal pricing 节电电价法Payoff table 支付矩阵Property right 产权Public pricing 公共定价Quality regulation 质量监管Regional electricity market 区域电力市场Selling short 卖空Settlement account 结算帐户Supply-demand ratio 市场供需比Trading manner 交易方式Transaction cost 交易成本Whole energy competition 全电量竞争模式Unplanned outrage 非计划停运Variation margin 价格变动保证金Yearly regulation 年调节Zero-sum game 零和博弈Zonal pricing 区域电价法price elasticity of demand 需求的价格弹性No-load Cost 空载成本Quasi-fixed Cost 准固定成本Extra-marginal Production 边际外生产Deadweight Loss 无谓损失Option Markets 期权市场Swap Contracts 互换合约Option Valuation 期权估值Exotic Options 特种期权Economies of Scale 规模经济CPF continuation power flow 连续潮流PLF probabilistic load flow 概率潮流Reserved capacity 预留容量ETC existing transmission commitment. 现有输电协议forecasting time span 预测时间跨度PM Probabilistic margin 概率界限EDS economic dispatch system 经济调度系统21。
无线局域网概览无线局域网(WLAN )一般是一个有线局域网的延伸。
无线局域网组件将数据包转换成无线电波或红外线(IR )光脉冲,并传送到其他无线装置,或到一个作为通往有线局域网效劳的接入点。
现今大多数无线局域网是基于IEEE 和标准之间的无线通信设备和网络。
这些标准予诺数据传输别离在1到2 Mbps 或5至11 Mbps。
并指定一个一起的架构,传输方式,及其他方面的无线数据传输,以改善互操作性的产品。
技术在设计一个无线局域网解决方案时,制成品无线局域网有多种技术选择。
每种技术来都有自身的优势和局限性。
窄带技术窄带无线电系统在一个特定的无线电频率传输和接收用户信息。
窄带无线电频率维持其频带为可能方才通过信息的大小。
在不同的频道频率认真和谐不同用户可幸免通信渠道之间产生不良的串音。
私人线很像一个电台频率。
当每一个家庭在一个居委会有自己的私人线,人们在一个主页不能听其他家的。
在无线电系统,隐私和不干与是完成所要求作出其他的家园。
无线电接收器过滤掉所有的无线电信号,除那些指定的频率。
从客户角度来看,有一个缺点,窄带技术是最终用户,针对每一个招聘网站的地址,必需取得FCC的许可。
扩频技术大多数无线局域网系统利用扩频技术,军事开发的宽带射频技术使用在靠得住,平安,关键任务通信系统。
扩展频谱是旨在提供高的带宽利用率,靠得住性,完整性和平安性。
换言之,在窄带传输中更多的带宽被消耗,但衡量所产生的信号,在成效,声音方面,更易侦测,只要接收机知悉对扩频信号播出的参数。
若是接收机没有调整在正确的频率,扩频信号看起来像背景噪音。
有两种类型的无线扩频:跳频和直接序列。
跳频扩频技术跳频扩谱(fhss )不管发射机和接收机,在一个众所周知的模式利用窄频载波转变的频率。
妥帖同步,净成效是要维持一个单一的逻辑通道。
关于一个意想不到的接收器,fhss似乎是短时脉冲噪声。
直接序列扩频技术直接序列扩展频谱(扩频)为每个位元生成一个多余的比特模式,以转交。
780-440-6737Edmonton 780-466-2362 Canada T6N 0A7 E-mail: josephwang@Memory GaugeDDIC SeriesGENERALGolden Canada Ltd. developed DDI-Cseries to provide an alternative to thehigh power tools on the market today.With the DDI -C series you can runlong tests without having to stackbatteries to get the test in. Our toolswill fill the memory with any samplerate on a single lithium “AA” battery.The savings in battery costs alonemake these tools a viable alternativeto other tools on the market today.Golden Canada Ltd. has been able to maintain high quality data and fast pressure/temperature response with one battery. Low power was not our only goal in building a memory recorder. We have also made every attempt to build a user friendly, durable and dependable downhole tool. We use one software package and one interface box to program and download all of our tools. All housings are made of 718 INCONEL orequivalent material.The DDI-C series come in a variety of pressureranges from 750 psi to 15000 psi. We also carry afast response temperature tool in both 1.25” and3/4”. The sample rate can be set as low as 1sample per second. With 696000 sample the toolwill run for 4 days on a 1 second rate. The memoryhad been doubled to 1380000 samples. Even withthe memory doubled the recorder will still fill thememory with any sample rate on a single “AA”battery.Memory GaugeDDIC SeriesSPECIFICATIONSTransducer Type...........................................................................Silicon Crystal Pressure Accuracy.................................................................. ..0.02% Full Scale Pressure Resolution...............................................................0.0003% Full Scale Pressure Drift................................................................................<3 psi / year Pressure Rate.............................................................. ...6000, 10000 ,15000psi Temperature Accuracy.........................................................................±0.20 °C Temperature Repiatability……………………………………………………………………………..…<±0.02℃Temperature Resolution.........................................................<0.001℃Full Scale Temperature Rate.........................................................................125°C, 150°C Power Requirements (volts)..................................................................3.3 V DC Power Requirements (battery life).................................... 1 AA Lithium Cell / year Memory Capacity............................................................1,380,000 Sets for time Sample Rate...........................................................1 sample / second to 18 hour Housing Material.............................................................718 INCONEL Standard Housing Diameter.....................................................................0.75 & 1.25 Inch Housing Length 0.75 Inch Recorder ..............................................12 & 20 Inches Housing Length 1.25 Inch Recorder......................................................22 Inches Communications...........................................................RS 232 / USB Comm. Box Software.............................................................................Windows Compatible Service. ...........................................................................................Sour(H2S)Continuous working time .......................................................................>90days Specifications subject to change without notice.ACCESSORIES•USB/RS-232 Serial Communication Box with Cable and 9V Battery•Pelican Case for Carrying Recorders and Accessories•O-rings•Battery Pack, “AA” Cell Lithium。
Energy Independence Is Not Energy SecurityMarch 28, 2006Bernard E. MunkBernard Munk, a Senior Fellow at the FPRI, is principal at Munk Advisory Services, an investment advisory service and publisher of ECOMENTARY, a private investment letter, and the website .The shock of Hurricanes Katrina and Rita in summer 2005, coming after a long run-up in petroleum prices stretching back from 2002, put energy independence back into the political arena. Unfortunately, discussions of energy policy, particularly by national politicians, offer an unlimited opportunity to pander and almost no thoughtful policy insights, particularly when the rubric, energy independence is the theme of the policy proposal.Energy independence—at least oil independence—is essentially unachievable in almost any environment that we can conceive over the next few decades. There is almost nothing we could do short of giving up our transportation-based economy that would secure this “goal.” To get there, we would have to abandon much of our passenger vehicular traffic use, airline availability, and the vast trucking network that makes this economy so flexible. The response of consumers to such an outcome is contradicted by the fact that despite much higher prices, consumption of transportation fuels has not fallen to a new, lower level that is likely to be sustained in the future.If talking about won’t make it happen, one ought to ask the next politician or policy wonk who starts from this premise, why are they taking up your time? It is misleading and counterproductive to make this appeal to voters.We use over 20.7 million barrels a day of crude oil and products and import 58.4% of that requirement. Looking at history, it is undeniab ly clear that our “dependence” is likely to grow, not diminish. The essential reality from which rational energy policy must start is that when it comes to transportation fuels (gasoline, diesel, and jet fuel), there are no suitable substitutes that pack as big a bang for their weight now available. Furthermore, we are unlikely to have a hydrogen economy suddenly appear out of a laboratory and be commercialized in any significant way over the next few decades. Can we replace oil in power generation or in space heating? Certainly. We have made nuclear technology much safer and far more efficient than the nukes built prior to the Three-Mile Island scare of 1979. We also have plenty of coal and the technology is available to make that coal a clean power source. While power costs are likely to rise, the U.S. requirements for power generation could be supplied domestically, even if U.S. natural gas reserves diminish in the future. If power generation was our only problem, the U.S. could be “energy independent.”Transportation is an entirely different issue. We need petroleum to run our transport system. We can make some improvements in miles per gallon; we can, at higher cost, substitute some addition to our burnable fuel supply produced from agriculture, and we co uld use more natural gas (again, if the price didn’t matter) for various kinds of “fleet operations” (e.g. municipal buses). But unless the American public is prepared to accept a very different kind of transportation economy, energy independence issimply a fiction. It is not the basis for a sound energy policy for the United States. Isn’t it time to re-frame the issue?Voters should understand these facts and question politicians who claim differently. High prices are not our problem, but the can be our friend. Higher prices for crude oil and gasoline will allow the economy to adapt to the new reality that the costs of finding new sources of crude oil for a growing world economy will certainly rise. We have a very flexible economy with lots of opportunitie s to “substitute,” and higher prices will amplify those opportunities by reducing the demand for oil and gas products while inducing more innovation and higher productivity in the search for and refining and distribution of oil and gas products.Energy security, however, is important, and to a large extent achievable, but it should never be confused with energy independence. We can improve the former but we cannot ever have the latter. We are unlikely to eliminate hurricanes even if it turns out we learn to better predict their path and intensity. Putting 20 to 25% of our national refinery capacity in a likely path for hurricanes is a risk we have undertaken. They can be damaged even if our weather predictive capacity is sharply improved. Further, we can lose supplies of crude oil coming from the Gulf of Mexico. In the tightly supplied world of crude oil, the loss of 1.5 million barrels of oil from the world supply is significant. We will have to learn to live with these choices or build refineries in areas less susceptible to weather disturbances.Uncertainty also extends to the likely disturbances in the international political economy that will affect energy supplies. Much as we might like, the political fortunes of oil suppliers are beyond our practical control. Upheavals can and do take place. More significantly, as the Third World becomes the Second and First World, the demand for energy in many countries is going to rise. Those demands will entail rising energy prices and a more active market in the search for and control of foreign energy supplies. That said, however, energy suppliers need markets and the “oil weapon” is much less likely to be used by “rogue” regimes for the very fact that other than oil or natural gas, those regimes generate few alternative foreign exchange earners. Their inhospitable attitudes toward foreign investors, particularly the United States, more often than not, confines them to the so-called “oil curse” and the attendant corruption that makes for permanent poverty.It is time for national political leaders to realize that a call for energy independence is an atavistic appeal to an independence that can never be. It is time to learn to live with an energy world that is, not the world that can never be!A good analogy was the early 1960s, when we were in the midst of the Cold War and both sides were armed with nuclear weapons. Our building nuclear weapons did not enhance our feeling of security. Conversely, the Russians could build more of them and still feel insecure. Remember the 1964 film, Dr Strangelove. It was subtitled ’How I Learned to Stop Worrying and Love the Bomb.’ Both sides accepted that we lived in an uncertain world. We built massive offensive capability to deliver weapons against the Russians, but it didn’t bring us true security. The film satirized the circumstances of our insecurity when Major T.J. “King” Kong (Slim Pickens) rode the bomb out of our rogue bomber yippy-i-aying like it was a bucking bronco asthe weapon plummeted to Earth. General Turgidson certainly showed the “Russkies” that they couldn’t push the U.S. around!Oil is similar. We will always have a certain amount of “insecurity” owing to the nature of how oil is “owned.” Energy resources are more often than not owned by States who react to political themes not necessarily economic incentives. Recently, the WSJ reported that less that only some 7% of the world’s energy resources are ’in play’ in the market while the rest are either directly or indirectly controlled by States. That means that energy security can only come from political cooperation enhanced by economic incentives.Several countries who are important oil producers currently have governments who don’t like us (Iran, Venezuela, and possibly Russia). However, they need the revenues from oil and gas sales and whether the U.S or other countries receive buys them, it is the total pool of hydrocarbon resources that bears on prices and therefore on our economy. Whether we ’control’ them may have little to do with how they are priced. Perhaps the m ost important lesson of “Katrita” was how quickly the International Energy Agency (IEA) made supplies available to us that helped to dampen our “insecurity.” That should have been a signal lesson from the catastrophe. Instead, we are in a blame game over how well prepared we were and who fumbled the ball.We should forget about independence and concentrate on measures to improve our energy security, starting with recognizing the mutuality of interests among consuming and producing countries and building our policies around improving our relationships with other consuming and producing countries. All of these countries have a mutual interest in a world that has as few “shocks” in it as possible.Given that the State is so fundamental to the energy market, and that oil is a global business with every country having an interest in that the oil market is not frequently upset and interrupted, energy security is just one aspect of international security. We have long ago learned that we have to find common paths along which we can cooperate with other countries rather than finding all the elements that create a competitive environment between States. I can give a number of examples. Consider the recent case of the Chinese National Oil Company, CNOOC, which wanted to buy Unocal, a holder of energy concessions in other countries. Congress and the White House created a furor and blocked the sale to CNOOC. But look at the real damage such and intervention provided. What we should be caring about is to increase the total supply of oil made available to the world market. We don’t require that oil be produced by an American firm. We want our investments in China to be treated on a non-discriminatory basis, but we sent a powerful signal to the Chinese that we don’t want to pl ay by these rules—-particularly when a ’vital national interest’ is at stake! We intervened on behalf of a U.S. company, probably reducing the price of the shares (to American stockholders) as well. We antagonized the Chinese and told them play by our rules except when we want to play by different ones. Is that the signal we want to send to the world? Is this the cooperative environment in which international security can be enhanced?A legacy of that intervention is the forced “give-up” by a U.A.E.-based company (Dubai Ports World) of the management of six U.S. ports. The furor was over“security,” although there has been no evidence produced that the “ownership” as opposed to the “conduct” or “management” of the ports would result in less security for the U.S. What has been demonstrated, however, is that this country is not “rule-based,” but is as jingoistic as many others. We have descended into a form of protectionism that can only destroy international security.Under almost any conceivable circumstances, Dubai Ports would have been much more alert to the dangers of terrorism than any other non-Arab company. Why? Precisely because the entire world would be watching their conduct. We need to promote the idea of common interests in security in Arab countries. Throwing down the gauntlet of “national interest” was hardly the way to promote international cooperation.What about high-priced oil? There is good news and bad news here, but more good news than bad news. The bad news is obvious: it will cost more to fill up your SUV than it used to. The good news is that the return to investment in petroleum discovery and refinery has gone up. High prices send a signal to investors to deploy more resources into energy development, and at the same time signal consumers to slow down domestic consumption. Both activities improve the margin of safety of petroleum supplies and that is in all of our interests.What the political leaders of this country need to understand is why do we have high prices anyway? Chronic under investment in the energy business and sustained, world economic growth. If they want to blame someone, blame it on Alan Greenspan. He promoted world economic growth like no other Fed Chairman before him, allowing the U.S. to have one of its longest, sustaining periods of economic growth. Our growth set off rapidly rising petroleum consumption at home and high growth in many developing countries. By elevating their growth, they elevated their demands for petroleum. Welcome to a “flat world!”What the Congress has done recently is go backwards on the world’s progress toward interdependence and the U.S. commitment to the free flow of goods and capital. That is precisely the wrong signal for international cooperation. The Congress is patting itself on the back that they have stood up to globalization and are more vigilant about U.S. security than is the President. They should be ashamed because they are signaling the world that the U.S. has only a partial commitment to free trade and investment.The Siren Song of Energy Independence is as dangerous to us as the Sirens were to Ulysses. His trick should be ours! Bind up your ears, America, and pay attention to international security.SummaryEnergy independence is a false and misleading goal.“Energy independence is literally an impossibility in a globalized oil world. Globalization means interdependence. What happens over “there” has to affect us “here” and conversely. Everyone can predict the outcome of a physical interruption in oil shipments, but what happens when the U.S. economy grows spectacularly? We promote higher oil demands here and abroad!Energy security, however, is a worthwhile goal and we can improve ours by behavingresponsibly in world markets. Focus on free trade and the free movement of capital. Higher prices are the friend of energy security. Higher prices will induce more investment in energy production while restraining demand growth. That creates extra spare capacity in the world and a larger amount of spare capacity will produce more price stabi lity. What we don’t need is high variability in prices, because that creates uncertainty and uncertainty is the enemy of planned, rational and forward looking investment in the oil industryOther sources of variability in energy prices areInconsistent government policiesFrequent changes in taxes and incentives to drillFrequent changes in environmental policies that restrain refinery expansion International political disturbances that create uncertainty in futures markets which means higher prices.As Wa lt Kelly’s famous cartoon character Pogo put it, “I have seen the enemy and it is us!。
international business 国际商务economic integration经济一体化globalization of markets 市场全球化IMF 国际货币基金组织global financial system全球金融体系multinational company 跨国公司trade barrier 贸易壁垒GA TT 关税与贸易总协议market liberalization市场自由化emerging markets 新兴市场the globalization of capital资本全球化GNI国民总收入economic integration blocs经济一体化集团Coincide with 与......相符Foreign direct investment外国直接投资International portfolio investment国际证券投资Letter-head 信头Reference and date 编号与日期Inside name and address信内名称,地址Attention line 经办人Subject line 标题(事由)Complimentary close 结尾敬辞Carbon copy notation 抄送Identification line 辨认记号The second page第二页Trade with 经商,做买卖Trade in 经营(商品)deal indeal with 与...打交道,与...做生意,处理Commission house 提取佣金的商行Light industrial products 轻工产品State-operated corporation国营公司Joint venture 合资企业Deal exclusively in 独家经营Enjoy a high reputation享有很高的声誉Sister corporation 兄弟公司Supply sb. With sth.向某人供应某物Be assured of 使确信,使放心Get in touch with sb.与某人联系,接洽Do business with sb.与某人做贸易,成交Specific requirement 具体要求Through mutual efforts共同努力Equality and mutual benefit 平等互利In time 及时In turn 轮流地In short 总之Price list 价目单Cotton prints印花布Export terms 出口条款Fine in quality 质地优良Moderate in price 价格公道Novel in design 设计新颖Household electric appliance家用电器Energy-saving节能In order 妥当;整齐无误Make...to order 定做In good order 情况良好To one’s order以...为抬头,凭..指示Be subject to 以...为准By separate post 另邮Prompt shipment 即装Processing with supplied materials 来料加工Enquire for 询购某物Enquire about 打听某事Enquire into 调查Make sb an offer for 给某人报..的价格Sth be under offer 某商品在出售中The offer is firm for.....此报盘有效期为Make sb. an offer 向某人报盘Accept offer 接受报盘counter-offerRenew offer 恢复报盘还盘Decline offer 谢绝报盘Withdraw offer 撤销报盘Entertain offer 考虑报盘Non-firm offer 虚盘An offer without engagement 无约束性的报盘Wholesale price 批发价Gross price 毛价net price 净价Sell off 拍卖Sell out 卖完Repeat order 重复订购initial order初次订购Conditions of sales 销售条件A special allowance 特别折扣Subject to prior sale 以先售为条件Subject to change without notice 不经通知可以改变In cartons 装纸箱As per 按照Acknowledge receipt of 收到As a much lower price than 以比...低得多的价格Supply...from sock 供...现货Have...in stock 有...现货Out of stock 脱销With /in reference to... 关于For your reference 供你方参考Confirmed L/C 保兑信用证Commercial invoice 商业发票Make up 制作,准备Confirming bank 保兑行Confirming house 保付商行Shipping instructions 装船提示Agree on(upon)... 就...达成一致The rough handling 野蛮装卸Bill of exchange 汇票Promissory note 本票Demand draft(D/D) 票汇Telegraphic transfer(T/T)电汇Mail transfer(M/T)信汇Remitting/paying bank 汇出行/汇入行Collecting bank 托收银行Documents against payment 付款交单Documents against acceptance(D/A)承兑交单Commercial bill 商业汇票Banker’s bill 银行汇票Sight bill 即期汇票Time bill 远期汇票Irrevocable L/C 不可撤销信用证Unconfirmed L/C 非保兑信用证Non-transferable L/C 不可转让信用证Documentary L/C 跟单信用证Revolving L/C 循环信用证Reciprocal L/C 对开信用证Back to back L/C 对背信用证Red clause L/C 红条款信用证Standby letter of credit 备用信用证Clean L/C 光票信用证Reimbursing bank 偿付行Advising/opening bank 通知行/开证行Negotiating bank 议付行Give commission on.....对...商品给佣金At a discount 打折In payment of 支付In duplicate/triplicate一式两份/三份Draw up a contract 起草合同draft a..Have a contract ready for signature备好合同签字Sign(close,enter into) a contract 签订..Execute a contract 履行合同U31.我们在平等互利的基础上与其他国家进行贸易。
Energy Market Disruption StrategiesThe energy market is undergoing significant changes and disruptions due to advancements in technology, evolving consumer preferences, and increasing environmental concerns. In this article, we will explore some of the strategies that companies can employ to navigate and thrive in this rapidly changing landscape.1. Embrace renewable energy sources: One of the most effective strategies for disrupting the traditional energy market is to embrace renewable energy sources such as solar, wind, and hydroelectric power. By investing in renewable energy technologies, companies can reduce their reliance on fossil fuels, lower their carbon footprint, and appeal to environmentally-conscious consumers. This shift towards renewable energy not only helps companies stay ahead of regulatory changes but also positions them as leaders in sustainability.2. Implement advanced grid technologies: Another key strategy for disrupting the energy market is to implement advanced grid technologies such as smart meters, energy storage systems, and demand response programs. These technologies enable companies to better manage energy supply and demand, optimize grid efficiency, and reduce operational costs. By leveraging data analytics and automation, companies can improve grid reliability, resiliency, and flexibility, thereby creating new opportunities for innovation and growth.3. Foster collaboration and partnerships: In an increasingly interconnected energy market, collaboration and partnerships play a crucial role in driving disruption and innovation. Companies can collaborate with energy regulators, policymakers, utilities, and technology providers to develop integrated solutions, pilot new business models, and overcome market barriers. By fostering an ecosystem of cooperation and knowledge sharing, companies can accelerate the adoption of disruptive technologies and create value for all stakeholders.4. Embrace digital transformation: Digital transformation is a powerful strategy for disrupting the energy market and unlocking new revenue streams. Companies canleverage digital technologies such as artificial intelligence, Internet of Things, and blockchain to optimize energy production, distribution, and consumption. By utilizing real-time data analytics and predictive modeling, companies can enhance operational efficiency, improve customer engagement, and create personalized energy services. Embracing digital transformation enables companies to stay competitive in a rapidly evolving market and deliver innovative solutions to meet changing customer needs.5. Engage with stakeholders and communities: Lastly, companies should focus on engaging with stakeholders and communities to build trust, transparency, and support for their disruptive energy strategies. By communicating effectively with customers, employees, investors, and local communities, companies can demonstrate their commitment to sustainability, social responsibility, and long-term value creation. Engaging with stakeholders fosters mutual understanding, enhances reputation, and builds a strong foundation for successful energy market disruption.In conclusion, the energy market is ripe for disruption, and companies that proactively embrace change, innovation, and collaboration will thrive in this dynamic environment. By embracing renewable energy sources, implementing advanced grid technologies, fostering collaboration and partnerships, embracing digital transformation, and engaging with stakeholders and communities, companies can position themselves as leaders in the evolving energy market. By adopting these strategies, companies can drive positive social and environmental impact, achieve long-term profitability, and create a sustainable energy future for generations to come.。
The Energy Charter Treaty is an international agreement that aims to promote the liberalization of energy markets and ensure the security of energy supply. The treaty was adopted in 1994 and has been signed by over 50 countries, including major energy producers and consumers such as Russia, Saudi Arabia, France, and the United States.The Energy Charter Treaty establishes a framework for cooperation among its signatories in the areas of energy trade, investment, and transit. It also includes provisions for dispute resolution and the protection of investors' rights.Some of the key provisions of the Energy Charter Treaty include:1. Freedom of energy supply: Signatories are committed to ensuring the freedom of energy supply, including the right to access foreign markets for their energy resources.2. Investment protection: The treaty provides for the protection of investments made by signatories in each other's energy sectors, including those related to exploration, production, transmission, distribution, and storage.3. Transit rights: Signatories are obligated to ensure the safe and uninterrupted transit of energy through their territories, including pipelines and other infrastructure.4. Dispute resolution: The treaty provides for a dispute resolution mechanism, known as the International Centre for Settlement of Investment Disputes (ICSID), which can be used to resolve disputes between signatories.Overall, the Energy Charter Treaty is an important tool for promoting energy security and market liberalization around the world.。
储能溢价能源交易英语"Energy storage premium and energy trading" in EnglishThe term "energy storage premium" refers to the additional cost or value associated with energy storage. Energy storage systems, such as batteries, can store energy generated from renewable sources, like solar or wind, and release it when needed. The premium attached to energy storage reflects the added benefits it brings, including the ability to manage energy supply and demand, enhance grid stability, and enable the integration of intermittent renewable energy.On the other hand, "energy trading" refers to the buying and selling of energy, typically electricity or other forms of energy, on a market. Energy trading involves the exchange of energy between suppliers and consumers, and it can occur at various levels, including wholesale markets and retail markets. Energy trading aims to optimize energy utilization, ensure energy supply security, and facilitate the efficient functioning of the energy market.Together, energy storage premiums and energy trading play important roles in the energy industry. Energy storage premiums reflect the value of storing energy and the associated costs, while energy trading enables the buying and selling of energy to meet demand and manage supply. These concepts are crucial for the effective operation and development of the energy market, as they help to balance supply and demand, promote the integration of renewable energy sources, and ensure the stability and reliability of the energy system.。
Energy-Security Tradeoff in a Secure Cache Architecture Against Buffer Overflow AttacksKoji Inoue† §†Department of Informatics, Kyushu University. 6-1 Kasuga-Koen, Kasuga, Fukuoka 816-8580 Japan §PRESTO, Japan Science and Technology Agency, 4-1-8 Honcho Kawaguchi, Saitama 332-0012 Japaninoue@c.csce.kyushu-u.ac.jpABSTRACTIn this paper, we propose a cache architecture, called SCache, to detect buffer-overflow attacks at run time. Furthermore, the energy-security efficiency of SCache is discussed. SCache generates replica cache lines on each return-address store, and compares the original value loaded from the memory stack to the replica one on the corresponding return-address load. The number and the placement policy of the replica line strongly affect both energy and vulnerability. In our evaluation, it is observed that SCache can protect more than 99.3% of return-address loads from buffer-overflow attacks, while it increases total cache energy consumption by about 23%, compared to a well-known lowpower cache.1. INTRODUCTIONAs the popularity of mobile computing devices and the advance in internet information services, considering energy-security efficiency of computer systems becomes more important. Although the internet is a much useful instrument, it also gives an opportunity for attacking remote connected devices to malicious persons. On the other hand, reducing energy consumption is an inevitable design constraint for mobile devices such as laptop computers and cellular phones, because it affects directly the battery lifetime. Fundamentally, in order to defend own computer system, some extra tasks which do not contribute program-execution results are required, resulting in wasted energy consumption. However, only few attempts have so far been made at the tradeoff between energy and security. In this paper, we propose a cache architecture, called Secure Cache (SCache), to prevent buffer-overflow attacks. Moreover, we evaluate the energy overhead caused by SCache based on a 0.18µm SRAM design, and discuss the tradeoff between energy and security. The attackers attempt to alter the procedure return address by causing a buffer overflow which breaks the structure of memory stack. SCache detects thereturn-address corruption without any software supports at run time. When a return address is pushed onto the memory stack, SCache generates one or more replicas of the return-address value, and saves them into the other cache space. Since on-chip caches give a large impact on both performance and energy consumption, researches have proposed a number of approaches to lowering cache energy [6][7][8][10]. In contrast to prior work, this paper focuses on the tradeoff between energy and security. This paper is organized as follows: Section 2 explains briefly the mechanism of buffer overflow, and introduces related work to solve the bufferoverflow problem. Section 3 proposes SCache architecture, and the organization and operation are explained in detail. In Section 4, we evaluate vulnerability and energy consumption of several SCache models, and discuss the energy-security tradeoff. Finally, in Section 5, we conclude this paper.2. STACK SMASHING ATTACKIn this section, we explain the mechanism of the stack smashing, and show related work to prevent the buffer-overflow attack.2.1. Buffer-Overflow VulnerabilityTo attack vulnerable computer systems, at least two processes have to be done: injecting an attack code and hijacking the program-execution control. The buffer overflow makes it possible to achieve both the issues simultaneously. This is one of the main reasons why the buffer-overflow vulnerability is commonly exploited for attacking target computers. For example, the malicious programs such as Code Red warm in 2001 and Braster in 2003, which raged in the world, utilize this defect. Figure 1 illustrates the percentage of CERT advisories relating to the buffer-overflowCERT Advisories relating to buffer-overflow (%)60 50 40 30 20 10 0 1996 1997 1998 1999 2000 2001 yearHigher Addr. FPStack Growths Return Addr. Saved FPFPs Corrupted Ret. Addr.int f ( ) { … g (s); … } int g ( char *s) { char buf [10]; … strcpy(buf, s); } Executed code exampleSP Lower Before the Addr. attackLocal String Variables Growth (buf) SPAttack CodeStack SmashingFigure 1: buffer overflow advisories (from [9]) vulnerability for each year. We see from Figure 1 that 50% of advisories in 2001 relate to this weak point. The buffer overflow is caused by writing an inordinately large amount of data into a buffer. Unfortunately, the C programming language does not perform automatically array-bound checks, and this defect mainly exists in the standard C library such as strcpy(). Therefore, many programs have the possibility to suffer from the buffer-overflow vulnerability. The overflow breaks memory stack structure as depicted in Figure 2, and this operation is called stack smashing. In this figure, we assume that the function f() calls the function g(), which includes a vulnerable operation strcpy(), as shown in the rightmost figure. The state of the memory stack immediately after the function call g() is depicted in the leftmost figure. The stack consists of the function parameters, the return address to the caller, the previous frame pointer, and the local array variable buf. In g(), if the size of the string pointed by s is larger than the memory size allocated for the local variable buf, a buffer overflow takes place when the strcpy() function is executed. As a result, the contiguous stack contents are overwritten. If the pointer s points a malicious string which is meticulously constructed by an attacker, the attack code is injected into the memory stack and the return address is altered to the top of the injected code, as shown in the middle one in Figure 2. The corrupted return address is set to the program counter (PC) when the execution of g() completes. As a result, the program-execution control is finally hijacked by the injected attack code.Figure 2: Stack Smashing2.2. Related WorkThe most straightforward way to solve the bufferoverflow attack is to prohibit the execution of codes stored in data segments. For example, AMD Athlon64 employs this protection. However, some programs attempt to generate an executable code at run time, e.g. just-in-time compiler. To support such operations, the microprocessors should be able to execute the instructions stored in a data segment. So far, many techniques to address the bufferoverflow attacks have been proposed. They can be classified into two types: static and dynamic. The static approach generates a secure object code based on source code analysis. For example, the paper [11] formulates detection of buffer overflows as an integer range analysis problem in order to find the potential of stack smashing. SASI introduced in [4] inserts reference-monitor codes into application programs to observe program-execution behavior. For instance, a bound-checking code as a reference monitor may detect buffer-overflow attacks. StackGuard, which is a patch to gcc, is another static approach to defending the stack smashing[3]. Each return address is pushed onto the stack with a “canary word” which is a randomly generated value. The canary is allocated to the next stack entry of the return address, and its copy is also stored into a general-purpose register. Therefore, we can detect stack smashing by comparing the canary value read from the stack memory with that saved in the register. This approach stands on the assumption that the canary word is altered whenever a return address corruption takes place. One of the main drawbacks of the static approach is code compatibility, because it requires a code translation or re-compilation. On the other hand, thedynamic approach does not have this negative effect. A dynamically loadable library (DLL) called libsafe is used to check vulnerabilities before un-trusted standard C libraries are executed[1]. Similarly, libverify is another DLL which injects a verification code at the start of the process execution via a binary re-writing[1]. StackGhost exploits cleverly the register window of the SPARC architecture[5]. In such architecture, a return address is saved in the register window. Only when a register-window overflow occurs, the return address is stored into a memory stack by OS support. Although the memory stack access is potentially vulnerable, OS can insert some operation to protect the return address, e.g. implementing a canary word as well as StackGuard. Another approach called SRAS (Secure Return Address Stack) is an LIFO fashion small memory embedded in the microprocessor core[9]. SRAS is a straightforward but very efficient architectural support to prevent the buffer-overflow attacks. A return address is pushed onto not only the memory stack but also SRAS, and they are compared to detect stack smashing when the corresponding return instruction is executed. Our research has been started from this paper. Since SCache belongs to the dynamic approach, code compatibility can be maintained. In addition, we do not modify the library and OS. Unlike StackGhost, our approach is independent of the microarchitecture of processor. We exploit the random-access large cache to store the copied data. Therefore, SCache has enough capacity, and can work well even if function calls are performed with non-LIFO fashion, e.g. longjump(). Another simple hardware support is to XOR the return-address value with a secret key, whereas SCache does not require any key information. Furthermore, the main difference of this paper is to focus on the energy-security tradeoff that is hardly discussed so far.RL: Replica Line ML: Master Line R-flag Tag Ref. Addr. Tag Offset Tag Match && R-flag Tag Match && no R-flag HIT? IndexData (Ret. Addr.) way0Store (push)Line Original Replica way1 way2 way3 ML master 32 32 RL RL replica replica Replica-MUX Master-MUX32-bit comparator Safe?Data (Ret. Addr.)Load (pop)Figure 3: 4-way set-associative SCache order to achieve such a hardware protection, SCache attempts to make one or more copies of the return address when it is stored into the cache. We call a cache line including the copy of the return address a replica line. Actually, the replica line is not a complete copy of whole cache line. Only the return address is copied. Since the replica line can be placed in the cache set indexed by the reference address of the current return-address store, we can generate at most Asso-1 replica lines where Asso is the cache associativity. Any cache accesses, except returnaddress stores, can not overwrite the replica lines. Therefore, the replica lines are treated as read only. When a return address is popped off the memory stack, SCache selects one of the replica data and compares it with the popped original return address. If they are exactly the same, we can ensure that the popped return address is safe. Otherwise, it means that a return-address corruption takes place, thus a signal to report security status is sent to the processor in order to terminate the current program execution.3.2 Structure and OperationFigure 3 shows the structure of a four-way setassociative SCache. Here, we assume that the number of replica lines to be generated for each return-address store (noted as Nrep) is two. Unlike conventional caches, one-bit replica flag (R-flag) is added to each tag entry. The R-flag is set to one if the corresponding cache line is a replica, otherwise it is reset to zero. Moreover, a multiplexer to select a replica data and a 32-bit comparator for examining return-address corruption are required. Figure 4 illustrates its3. THE SCACHE ARCHITECTUREIn this section, we propose the SCache architecture, and explain the structure and operation of a four-way set-associative SCache.3.1 OverviewCommonly, return addresses are transferred to (or from) the memory stack through on-chip caches. Therefore, if it is possible to protect the return address on the cache, we can prevent stack smashing without affecting the structure of complex microprocessor. Inoperation on return-address load/store hits. Whena return address is pushed, the cache works as follows.W1. With the same manner as conventional caches, the set indexed by the reference address is accessed for tag checking. The corresponding Rflags are also read in parallel. W2. Then the return address is stored into the cachehit line. We call the target line the master line. Namely, the master line includes the original value of the return address. W3. If some replica lines already exist in the indexed set, the return address is stored into them in order to avoid a coherence problem. The replica lines can be detected by examining the tag and the R-flag at the step W1, because the corresponding replica lines have the same tag information and a valid R-flag. W4. New replica lines are generated by writing the return-address value into other non-replica line(s) and setting the corresponding R-flag(s) to one. The tag and offset information of the original write operation performed at the step W2 is used to make the replica line(s). Note that the example assumes a cache hit. However, even on a miss, the cache works in the same manner after the line replacement is completed. As explained in Section 3.1, SCache does not make a complete copy of whole master line, but just writes one word return-address value to the several lines in parallel. In other words, the contents of the replica lines are not exactly the same as those of the master line. In addition, a replica line can include several return addresses if the offset addresses are different. On the other hand, at a return-address load, the cache examines whether or not the popped return address is safe as follows. R1. The cache activates all the ways to read cache lines, tags, and R-flags. Then, the master line, which has the matching tag and the invalid Rflag, is selected. The return-address data read from the master line is sent to the processor. R2. The replica lines having the matching tag are searched in the referenced set. If there are several replica lines corresponding to the current return-address load, one of them is randomly selected. Otherwise, the cache reports to the processor that the loaded return address may be unsafe, and completes the current cache access.Cache Access (for Return Addr.)RL: Replica Line ML: Master LineStoreyesStore?noLoadStore the Return Addr. into ML Store the Return Addr. into existing RLs (if it has the same tag) Generate RLs until #RL == NrepRead lines from all ways (provide the data to CPU) RL Hit?noyesReturn Addr. Checkun-matchNormal Complete Error!matchSafe CompleteUnsafe CompleteFigure 4: Operation on return-address accesses R3. The return address read from the master line is compared with that obtained from the selected replica line. For this examination, only one replica data is selected and compared with the original one, because all of the replica lines include the same copy of the original returnaddress value. If the 32-bit comparison result is not a match, the cache alarms to terminate the program execution due to the insecure returnaddress load. In conventional write accesses, the tag checks and the data write (W1 and W2) are performed sequentially. Although SCache requires two more steps, W3 and W4, they can be executed in parallel with the step W2. For read accesses, SCache completes to read the target data at the step R1 as well as conventional caches. In addition, the hardware components for SCache, a replica-line MUX and a 32-bit word comparator, do not affect the cache critical paths. Accordingly, SCache does not worsen the cache-access time. For normal accesses (i.e. non-return-address loads or stores), SCache operates as the same as conventional caches except that the invalid R-flag is included in the cache-hit condition. Therefore, the normal accesses do not modify the replica lines. For instance, let us consider a write operation to the next stack entry of the return address. Here, we assume that the write target entry is allocated to the same cache line as the return-address value. In this scenario, the write operation is performed only to the master line, which has the invalid R-flag. The master line does not have any protection mechanism, thus the write operation can be completed.The microprocessor needs to output a signal to indicate whether the current access targets a return address. This can be easily achieved by checking the source (or destination) operand of the current memory reference [9]. For many microprocessors, the return address is located to a special register, e.g. R31.algorithm for replica-line placements. However, they affect negatively cache energy consumption and miss rates.4. EVALUATIONIn this section, we define six SCache models, and discuss the ability of proposed approach. The SCache models differ in the number and the placement policy of replica lines. First, we explain how the energy and vulnerability are evaluated. Next, we measure them for the SCache models, and discuss the energysecurity tradeoff. The performance overhead caused by SCache is also evaluated.3.3 Impact on Energy and SecuritySCache attempts to protect return addresses by generating replica lines. However, for each returnaddress store, the cache needs to write the returnaddress value into several locations, i.e. one to the master line and one to each replica line, thereby increasing write-access energy. Furthermore, on return-address loads, the cache examines all the ways to find replica lines. Namely, SCache can not reduce read-access energy as well as already proposed lowpower caches which attempt to avoid unnecessary way activation [6][10]. Another drawback of SCache is the energy overhead caused by degrading cache-hit rates. Since generating replica lines pollutes effective cache area, the energy consumed for cache-line replacements is increased. On the other hand, from the security point of view, SCache can detect return-address corruption whenever at least one replica line exists. However, replica lines may be evicted from the cache, because they are also candidates for the cache-line replacement on misses. The most straightforward way to solve this issue is to prohibit evicting the replica lines from the cache. In this approach, we need to carefully release the replica lines with an appropriate timing. This is because a too-early-release degrades the efficiency of security, while a too-late-release pollutes unnecessarily the effective cache area. A simple strategy is to release the locked replica lines when the corresponding return-address load is issued to the cache. However, in this approach, still we have a possibility to take the too-early-releases. A number of microprocessors employ various predictions to gain performance, e.g. branch prediction, value prediction, etc. In such high-performance microprocessors, some instructions may be squished when a miss-prediction takes place. Therefore, the return-address load which is squished in later releases rashly the corresponding replica lines in the cache. Another approach to achieve secure operations is that we allow to evict the replica lines but try to make the replica lines reside in the cache as long as possible. We can consider at least two approaches: increasing the number of replica lines to be generated (Nrep) and employing the MRU4.1 Experimental SetupWe extended SimpleScalar tool set (ver.3.0d) [12] to support the SCache approach, and executed seven integer programs and four floating-point programs from the SPEC2000 benchmark suite [13]. The small input data set was used to complete the whole program execution. In this evaluation, we assume that the L1 data cache size is 16 KB, the line size is 32 B, and the associativity is 4. Furthermore, a 4-way superscalar out-of-order execution is assumed. For other parameters, we used the simplescalar default value defined in [2]. In this section, we refer to the issued return-address load (or store) to the cache as the IRA load (or store). We use the following equation to evaluate the efficiency for security, Vulnerability = (Nv-rald / Nrald) * 100, where Nrald and Nv-rald are the total number of IRA loads in the program execution and that of insecure IRA loads (i.e. return-address loads without any replica line), respectively. On the other hand, to evaluate the energy overhead, we use the energy model, Etotal = Erd + Ewt + Ewb + Emp, where Erd and Ewt are the total energy consumed for cache read and write accesses, respectively. Ewb is another energy overhead caused by evicting dirty lines from the cache due to replica placement. Emp is the energy dissipated for cache-line replacements. We assume that the energy consumed for a next-level memory access is ten times larger than that for an L1 cache read access. Based on a 0.18 µm CMOS technology, we designed a 4KB SRAM array and estimated energy consumption. The circuits have been optimized to meet 3.0 ns access time. After the layout, we have measured energy consumption by performingTable 1: Cache-Miss RatesModel Bench 164.gzip 175.vpr 176.gcc 181.mcf 197.parser 255.vortex 256.bzip 177.mesa 179.art 183.equake 188.ammp #IRA Load(Nrald) 4,930,467 5,627,709 37,519,156 992,419 45,466,527 22,101,265 18,147,017 4,727,396 32,466 3,580,827 6,307,839 CONV 5.22% 3.53% 4.26% 20.02% 4.13% 1.75% 2.31% 0.14% 42.93% 2.44% 36.27% LRU1-NR 5.23% 3.59% 6.06% 20.05% 4.25% 1.83% 2.31% 0.15% 42.93% 2.45% 36.29% LRU1 5.22% 3.56% 4.29% 20.02% 4.18% 1.79% 2.31% 0.15% 42.93% 2.44% 36.28% LRU2 5.22% 3.63% 4.37% 20.03% 4.44% 1.91% 2.32% 0.16% 42.93% 2.46% 36.31% MRU1 5.22% 3.59% 4.33% 20.05% 4.23% 1.82% 2.31% 0.15% 42.93% 2.45% 36.28% MRU2 5.23% 3.66% 4.43% 20.06% 4.55% 1.94% 2.32% 0.16% 42.93% 2.47% 36.30% ALL 5.25% 3.74% 4.64% 20.10% 5.07% 2.32% 2.45% 1.08% 42.93% 2.52% 36.38%IRA: Issued Return Address, CONV: Conventional6.1% 5.4%Hspice circuit simulations with extracted load capacitances. First, we obtained the energy for accessing 1-bit memory cell that includes sensing and pre-charging. Next, we calculated the average energy for each operation based on the total number of bits to be accessed. Then we multiplied the energy by the number of events occurred during the program execution. Since the cache associativity is assumed as four, we can generate at most three replica lines for each return-address store, i.e. Nrep = 3. Furthermore, there are two options for the replica line placement, LRU and MRU. Here, we define six SCache models: LRU1-NR, LRU1, LRU2, MRU1, MRU2, and ALL. LRU1 and LRU2 generate one and two replica line(s) with the LRU placement algorithm, respectively. MRU1 and MRU2 place replica line(s) on the MRU location except the master line. The ALL model makes the maximum number of replica lines. In these models, the replica lines are treated as the same as normal lines, i.e. they can be evicted from the cache. On the other hand, LRU1-NR prohibits evicting the replica lines, and they are released when the corresponding return-address load is issued to the cache. We compare the SCache models with a conventional low-power way-predicting cache, noted as CONV. This model attempts to activate only the hit way which includes the reference data by employing an MRU-base way prediction [6], thereby saving3.0% 2.5%Vulnerability31.1%8.7%4.7%2.0% 1.5% 1.0% 0.5% 0.0%17 5.v pr 25 5.v ort ex 16 4.g zip 17 6.g cc 18 1.m cfLRU1-NR LRU1 LRU2 MRU1 MRU2 ALL25 6.b zip 217 9.a rt18 3.e qu ak e19 7.p ars er17 7.m es aBenchmarksFigure 5: Vulnerability cache-access energy. Here, we do not take the energy overhead caused by the way prediction, e.g. accessing an MRU table, into account. Note that the SCache models also perform the MRU-base way prediction. However, on each return-address load, they need to activate all the ways in spite of the correct wayprediction, as explained in Section 3.3.4.2 VulnerabilityFigure 5 shows the vulnerability of the SCache models. We should notice that conventional caches without any consideration for stack smashing have 100% of vulnerability. The number of IRA loads, Nrald, and cache-miss rates are also presented in Table 1.18 8.a mm p1.4Normalized Energy1.2 1 0.8 0.6 0.4 0.2 0x ke ip ip 2 cf a 5. vp 9. ar t c rs e rt e 6. gc 4. gz 1. m es 6. bz ua 7. pa 3. eq 16 18 17 8. am 5. vo 7. m 17 17 m p 18 r rEm p Ewb Ewt ErdCONV LRU1-NR LRU1 LRU2 MRU1 MRU2 ALL25192517Benchm arksFigure 6: Energy Consumption First, we discuss the effects of the number of replica lines, Nrep. In this simulation, we assumed that the LRU policy is employed for the cache-line replacement on misses. Therefore, replica lines generated in LRU1 are easily evicted from the cache due to conflicts. As we expected, LRU2 produces better results than LRU1 for all benchmarks due to the increased number of replica lines. We see the same situation for the MRU based SCaches, MRU1 and MRU2. If they can perform a strict MRU placement, the number of replica lines does not affect the vulnerability. However, when a return address is stored into the cache, all existing replica lines with the same tag information are updated in order to avoid coherence problem, as explained in Section 3.2. If a corresponding replica line already exists at the LRU location, MRU1 works as the same as LRU1. Therefore, MRU1 is more vulnerable than MRU2. For all but 181.mcf and 179.art, the most secure model ALL can protect more than 99.3% of IRA loads from stack smashing. Next, we discuss the impact of the replica-line placement algorithm. The MRU strategy constantly achieves higher security than the LRU models if the number of replica lines to be generated at each returnaddress store is the same. This is because the MRU placement makes the replica lifetime longer as well as increasing the number of replica lines. However, against our expectation, LRU1-NR does not work well for some benchmarks. One of the reasons of this result is a hasty release of replica data caused by squished return-address loads.4.3 Energy ConsumptionFigure 6 shows energy consumption and its breakdown for the SCache models. All results are normalized to CONV. From the figure, we see that increasing the number of replica lines worsens energy efficiency. The ALL model increases energy consumption by about 23% in the worst case, 197.parser. On the other hand, replica-line placement algorithm does not give a large impact on energy. The SCache models have the same energy overhead for read accesses. Since all of the ways in SCache are activated on each IRA load, the read-access energy depends on not the number of replica lines generated but the total number of IRA loads. In contrast, the energy dissipated for write accesses Ewt increases with the increase in the number of replica lines generated. This situation can be seen for Emp due to the increased cache-miss rates. For instance, in case of 177.mesa, the ALL model worsens the cache-miss rate from 0.14% to 1.08%, thereby increasing the energy for cache-line replacements Emp. For the programs with small energy overhead, 181.mcf and 179.art, it is observed that the increase in Erd, Ewt,181.5LRU22 1.5MRU12.4%MRU22.3%ALL1110.50.50.50164.gzip 175.vpr 176.gcc 197.parser 188.ammp0164.gzip 175.vpr 176.gcc 197.parser 188.ammp0164.gzip 175.vpr 176.gcc 197.parser 188.ammp(a) EVPand Emp are trivial. For these benchmarks, the total energy is originally dominated by Emp, 51% for 181.mcf and 62% for 179.art, due to the higher cachemiss rates as reported in Table 1. Therefore, the energy overhead caused by the replica lines is relatively hidden.(b) E2VPFigure 7: Energy-Security Tradeoff1.6% 1.4%Performance Overhead(c) EV2P1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0%LRU1-NR LRU1 LRU2 MRU1 MRU2 ALL4.4 Energy-Security TradeoffIn order to evaluate both energy and security at the same time, we introduce the energy-vulnerability product (EVP). We multiply the energy overhead by the number of vulnerable IRA loads. Furthermore, we measure the E2*V (or E*V2) product in order to consider more energy-oriented (or security-oriented) applications. Figure 7 shows the evaluation results for the five SCache models (LRU1-NR is not included). All the results are normalized to LRU1. From the figure, it is observed that each SCache model shows different characteristics. For instance, MRU1 produces the best performance if we see the energyoriented metric E2*V, while MRU2 or ALL give better results for security-oriented applications E*V2. This means that there is a tradeoff between energy and security, thus it is very important to explore the design space for coping both high security and low energy consumption. For the SCache approach, we conclude that MRU1 which increases the energy consumption only by 10% in the worst case but achieves relatively higher security should be selected if energy consumption is the primary design constraint. In contrast, MRU2 or ALL can protect more than 99% of IRA loads for many benchmarks, thus they are suitable for security-oriented applications.BenchmarksFigure 8: Performance4.5 Performance OverheadFinally, we evaluate the impact of the SCache approach on processor performance. As shown in Table 1, increasing the number of replica lines, Nrep, worsens cache-hit rates, thus the processor performance will be degraded. This negative effect appears clearly on the MRU-based models. Figure 8 reports the performance overhead caused by the SCache scheme. For the ALL model, the increase in execution time is at most 1.1%, 177.mesa. Furthermore, we see that in many cases the performance overhead is less than 0.4%. Therefore, we believe that the performance degradation caused by SCache is negligible.17 9. ar 18 t 3. eq ua ke 18 8. am m p16 4. gz ip 17 5. vp r 17 6. gc c 18 1. m cf 19 7. pa rs er 25 5. vo rte x 25 6. bz ip 2 17 7. m es a。