Podcast
Questions and Answers
What is the latency of a main memory reference for a double-word?
What is the latency of a main memory reference for a double-word?
What is the primary motivation behind building warehouse-scale computers using commodity components?
What is the primary motivation behind building warehouse-scale computers using commodity components?
What is the latency of sending 2 KiB of data over a 1 Gbps network?
What is the latency of sending 2 KiB of data over a 1 Gbps network?
Who predicted that computing would become a public utility?
Who predicted that computing would become a public utility?
Signup and view all the answers
What is the latency of reading 1 MiB sequentially from disk?
What is the latency of reading 1 MiB sequentially from disk?
Signup and view all the answers
What is the latency of a round trip within the same data center?
What is the latency of a round trip within the same data center?
Signup and view all the answers
What is the latency of sending a packet from California to the Netherlands and back to California?
What is the latency of sending a packet from California to the Netherlands and back to California?
Signup and view all the answers
What is the primary application of MapReduce programming in the context of warehouse-scale computing?
What is the primary application of MapReduce programming in the context of warehouse-scale computing?
Signup and view all the answers
What is the primary reason for the increase in latency from the local node to the array?
What is the primary reason for the increase in latency from the local node to the array?
Signup and view all the answers
What is the unit of measurement for bandwidth in Table 10.2?
What is the unit of measurement for bandwidth in Table 10.2?
Signup and view all the answers
What is the primary factor that collapses differences in bandwidth?
What is the primary factor that collapses differences in bandwidth?
Signup and view all the answers
What is the purpose of the core router in Fig. 10.3?
What is the purpose of the core router in Fig. 10.3?
Signup and view all the answers
What is the metric used to evaluate the efficiency of a WSC?
What is the metric used to evaluate the efficiency of a WSC?
Signup and view all the answers
What is the latency of DRAM in the local node according to Table 10.1?
What is the latency of DRAM in the local node according to Table 10.1?
Signup and view all the answers
What is the additional level in the networking hierarchy when a WSC needs more than one array?
What is the additional level in the networking hierarchy when a WSC needs more than one array?
Signup and view all the answers
What does the 'four nines' refer to in the context of Warehouse-scale Computers?
What does the 'four nines' refer to in the context of Warehouse-scale Computers?
Signup and view all the answers
What is the primary benefit of 'easy' parallelism, also known as request-level parallelism (RLP), in WSCs?
What is the primary benefit of 'easy' parallelism, also known as request-level parallelism (RLP), in WSCs?
Signup and view all the answers
What is a key factor in determining the location of a Warehouse-scale Computer (WSC)?
What is a key factor in determining the location of a Warehouse-scale Computer (WSC)?
Signup and view all the answers
What is the primary concern regarding operational costs in WSCs?
What is the primary concern regarding operational costs in WSCs?
Signup and view all the answers
What type of parallelism is crucial in WSCs, especially in web crawlers?
What type of parallelism is crucial in WSCs, especially in web crawlers?
Signup and view all the answers
What is the main benefit of using a hierarchical switch architecture in WSCs?
What is the main benefit of using a hierarchical switch architecture in WSCs?
Signup and view all the answers
What is a characteristic of massively parallel batch programs in WSCs?
What is a characteristic of massively parallel batch programs in WSCs?
Signup and view all the answers
What is the primary impact of the big scale of WSCs on their dependability?
What is the primary impact of the big scale of WSCs on their dependability?
Signup and view all the answers
What is the primary concern in terms of energy efficiency in Warehouse-scale Computers?
What is the primary concern in terms of energy efficiency in Warehouse-scale Computers?
Signup and view all the answers
What is the primary goal of dependability in Warehouse-scale Computers?
What is the primary goal of dependability in Warehouse-scale Computers?
Signup and view all the answers
What is the primary benefit of reducing costs in Warehouse-scale Computers?
What is the primary benefit of reducing costs in Warehouse-scale Computers?
Signup and view all the answers
What is the typical number of servers housed in a Warehouse-scale Computer?
What is the typical number of servers housed in a Warehouse-scale Computer?
Signup and view all the answers
What is the main goal of the networking equipment in a Warehouse-scale Computer?
What is the main goal of the networking equipment in a Warehouse-scale Computer?
Signup and view all the answers
What is the primary focus of Warehouse-scale Computers today?
What is the primary focus of Warehouse-scale Computers today?
Signup and view all the answers
What is the role of software redundancy in achieving dependability in Warehouse-scale Computers?
What is the role of software redundancy in achieving dependability in Warehouse-scale Computers?
Signup and view all the answers
What is the primary motivation for building Warehouse-scale Computers using commodity components?
What is the primary motivation for building Warehouse-scale Computers using commodity components?
Signup and view all the answers
What is the primary purpose of the Power Utilization Effectiveness (PUE) metric?
What is the primary purpose of the Power Utilization Effectiveness (PUE) metric?
Signup and view all the answers
What is the main reason for the increase in latency from the local node to the array?
What is the main reason for the increase in latency from the local node to the array?
Signup and view all the answers
What is the unit of measurement for latency in Table 10.1?
What is the unit of measurement for latency in Table 10.1?
Signup and view all the answers
What is the purpose of the core router in Fig. 10.3?
What is the purpose of the core router in Fig. 10.3?
Signup and view all the answers
Why do DRAM and Disk numbers are only available for the local node in Table 10.2?
Why do DRAM and Disk numbers are only available for the local node in Table 10.2?
Signup and view all the answers
What is the primary factor that collapses differences in bandwidth?
What is the primary factor that collapses differences in bandwidth?
Signup and view all the answers
What is the additional level in the networking hierarchy when a WSC needs more than one array?
What is the additional level in the networking hierarchy when a WSC needs more than one array?
Signup and view all the answers
What is the primary impact of the big scale of WSCs on their dependability?
What is the primary impact of the big scale of WSCs on their dependability?
Signup and view all the answers
What is the purpose of the map part in MapReduce?
What is the purpose of the map part in MapReduce?
Signup and view all the answers
What is the storage capacity of each node in the array?
What is the storage capacity of each node in the array?
Signup and view all the answers
What is the purpose of the reduce part in MapReduce?
What is the purpose of the reduce part in MapReduce?
Signup and view all the answers
What is the network bandwidth of each node in the array?
What is the network bandwidth of each node in the array?
Signup and view all the answers
How many nodes are there in each rack?
How many nodes are there in each rack?
Signup and view all the answers
What is the purpose of the EmitIntermediate function in MapReduce?
What is the purpose of the EmitIntermediate function in MapReduce?
Signup and view all the answers
How many racks are there in each array?
How many racks are there in each array?
Signup and view all the answers
What is the storage hierarchy of each computing node?
What is the storage hierarchy of each computing node?
Signup and view all the answers
Study Notes
Latency in WSC
- Latency in WSC increases due to networking software and switch overhead in the rack and array switch hardware/software.
- Latency numbers for DRAM, Flash, and Disk are:
- Local Node: 0.1µs (DRAM), 100µs (Flash), 10,000µs (Disk)
- Rack: 300µs (DRAM), 400µs (Flash), 11,000µs (Disk)
- Array: 500µs (DRAM), 600µs (Flash), 12,000µs (Disk)
Bandwidth in WSC
- Bandwidth numbers for DRAM, Flash, and Disk are:
- Local Node: 20,000 MiB/s (DRAM), 200 MiB/s (Disk)
- Rack: 1,000 MiB/s (Flash), 100 MiB/s (Flash)
- Array: 10 MiB/s (Flash)
Networking Hierarchy
- The network overhead increases latency and bandwidth differences between the local node and the array.
- In some cases, WSC needs more than one array, which adds an extra level to the networking hierarchy.
Power Utilization Effectiveness (PUE)
- PUE is a metric to evaluate the efficiency of a WSC.
- Example values for power usage: 0.5 (L1 cache reference), 5 (branch mispredict), 7 (L2 cache reference), 100 (main memory reference), 20,000 (send 2 KiB over 1 Gbps network), etc.
Cloud Computing
- John McCarthy predicted the concept of cloud computing as a public utility in 1961.
- Large Internet companies like Amazon, Google, and Microsoft use warehouse-scale computers to fulfill the increasing demand of users.
Warehouse-scale Computers (WSC)
- WSC needs to consider interactive and batch processing workloads, such as searches and social networking.
- Main requirements of WSC include:
- Ample parallelism (data-level and request-level parallelism)
- Operational costs (energy, power distribution, and cooling)
- Location (inexpensive electricity, proximity to Internet backbone optical fibers, and human resources)
- Big scale trade-offs (less dependability, bigger failure rates)
Nodes, Racks, and Switches
- WSC consists of nodes, racks, and switches, with an Ethernet switch on top of each rack and an array switch connecting multiple racks.
- Figure 10.1 illustrates the hierarchy of switches in WSC.
Warehouse-Scale Computers (WSC)
- A WSC is considered the foundation of Internet services used by billions of people daily, acting as one giant machine.
- It costs hundreds of millions of dollars for building, electrical and cooling infrastructure, servers, and networking equipment.
- The target is providing information technology for the world, instead of high-performance computing - HPC only for scientists and engineers.
Main Goals
- The main goals of WSC include the cost-performance relation, energy efficiency, and dependability.
- Cost-performance ratio: the work done per dollar is critical, with small cost reductions saving millions of dollars.
- Energy efficiency: the work done per joule is critical, with concerns about power and cooling.
- Dependability: achieved via redundancy, with a target of 99.99% availability, i.e., less than one hour of downtime per year.
Networking Hierarchy
- The network overhead considerably increases latency, and the network collapses differences in bandwidth.
- Some WSCs need more than one array, adding an extra level in the networking hierarchy.
- Regular "Layer 3" routers connect arrays together and to the Internet, with the core router operating in the Internet backbone.
Memory Hierarchy
- Each computing node contains:
- 16 GiB DRAM
- 128 GiB Flash
- 2,048 GiB Disk
- 1 Gbit/s Ethernet port
- The rack holds 80 nodes, and the array has 30 racks.
Latency and Bandwidth
- Latency numbers:
- DRAM: 0.1 µs (local node), 300 µs (rack), 500 µs (array)
- Flash: 100 µs (local node), 400 µs (rack), 600 µs (array)
- Disk: 10,000 µs (local node), 11,000 µs (rack), 12,000 µs (array)
- Bandwidth numbers:
- DRAM: 20,000 MiB/s (local node)
- Flash: 1,000 MiB/s (local node), 100 MiB/s (rack), 10 MiB/s (array)
- Disk: 200 MiB/s (local node)
Power Utilization Effectiveness (PUE)
- PUE is a metric to evaluate the efficiency of a WSC, given by Eq. (10.1).
Workload: MapReduce
- MapReduce is a popular framework for batch processing in WSC.
- The map part applies a programmer-supplied function to each input, producing an intermediate result of key-value pairs.
- The reduce part collects the output and collapses it using another programmer-defined function.
- Example: words and documents indexing using MapReduce.
Studying That Suits You
Use AI to generate personalized quizzes and flashcards to suit your learning preferences.
Related Documents
Description
This quiz covers the numbers of latency in microseconds for different storage devices and network components. It includes the overhead of networking software, switch, and array switch hardware/software.