Artificial Intelligence Authors: Elizabeth White, Zakia Bouachraoui, Liz McMillan, Pat Romanski, Carmen Gonzalez

RSS Feed Item

Thoughts on HPC, Grid and DataGrid

I work for a company called GigaSpaces , a company that specialises in providing software to enable the building of highly transactional, low-latency applications that need to be able scale-out and shrink back on demand.

Analysts tend to put GigaSpaces in a few different technology categories, namely Grid, DataGrid, Grid enabled Application Servers, or XTP (Extreme Transaction Processing). If you want a good concise overview of the current version of GigaSpaces XTP product, and what it can do I would suggest you check out the article by InfoQ and also the interview with Geva Perry , GigaSpaces chief marketing officer, at FishTrain.

There are some good screencasts which show how to download and setup the basic examples here and here . You can also choose to use GigaSpaces on Amazon EC2 - more about that here .

There have been a few times now where i've been with prospects who have no interest in the terminology other than its use for them in building a highly scalable application. To that end I've found myself explaining to them the differences / similarities between the various technology place holders, like Grid, DataGrid et al, that get bandied about.

I've done it often enough to think that it is worthwhile jotting this down here. There are my own thoughts and no way represent the opinions of GigaSpaces.

HPC is an acronym for ‘High Performance Computing’. This may seem to be stating the obvious, but HPC is starting to gather pace as an acronym for’ High Productivity Computing’ which brings together HPC, Grid, DataGrid, and virtualisation (of Middleware, storage and OS), so it worth knowing the difference.

High Performance Computing refers to computing systems and environments that typically use large numbers of processors, either as part of a single machine, or multiple computers that are organised in a cluster that operate as a single computing resource.

High-performance networking interconnects are used for Cluster based HPC systems. InfiniBand or Myrinet are an example of such interconnects. The network topology for such systems tends to be either a simple bus topology or (for extreme HPC requirements) a mesh topology. Mesh topologies tend to provide lower latency between each of the hosts. Networking performance and Transfer rates are therefore improved.

HPC business uses include data warehouses, line-of-business (LOB) applications and HPC provides the infrastructure for heavy transaction processing systems.

Grid computing, from a technology perspective, is still relatively new, hence why skillsets in this area are still valued at a premium.

Whereas with HPC you deploy a solution with a fixed number of nodes on dedicated hardware, Grid computing brings the flexibility of using standard non-hetregeneous hardware and standard operating systems, in which nodes can be added on demand. A standard software layers is applied across this infrastructure to achieve this. Key to the whole Grid proposition is that dedicated computing resources are not necessary, which leads to many Grids being built by reusing existing hardware to produce a powerful unified computing resource.

No special networking components are needed for Grid, and Grid computing is not limited to the local LAN. It is not unusual for MAN’s and WAN boundaries to be crossed when building a Grid resource.

A Grid can be thought of as a general computing resource in which different nodes of the Grid work, in parallel , on computational tasks that have been broken down by the Grid management software and farmed out using the Grid scheduler. Unlike traditional HPC solutions each node can be working on different computational algorithms and tasks which, after being farmed out, are brought back together to calculate the final result.

A good example of this in practice is in Investment bank where Grids are used to run Monte Carlo simulations to calculate risk With the adoption of Grid these type of calculations can now be run in near real-time allowing the banks to calculate risk positions on the fly which ultimately leads to shorter time to market, greater volumes and therefore increased revenue.

Perhaps an easier example to illustrate this is SETI , the Search for Extraterrestrial Intelligence. The SETI at home project searches for transmissions from ‘other’ civilizations by analyzing transmissions of cosmic origins to find patterns of communication. Clearly this is a vast undertaking and would require an infinite amount of compute resource. Cleverly the SETI project taps into the idle resources of millions of personal computers around the world, for the analysis of raw SETI data. This is done on by farming out work to home user PC’s that have installed a SETI screen saver. They analyze a chunk of work when idle and send the data result back, functioning very much like a vast on demand geographically disbursed Grid. Albeit one that has no latency constraints or end goal in sight !

DataGrid is a term bourne of out of the use of Grid technologies. In many ways it is a specialized used of Grid for certain types of applications that traditionally were not thought about when looking at applications to run on the Grid.

Grids are good for computationally intensive type of applications with fixed sizes of data to work on. These data elements form the input to the computation tasks and tend to be small. However Grids have been so successful in reducing TCO and ROA that it makes perfect sense to see what other type of applications they could be used for. Other application use cases tend to bring two types of issues that traditional Grid does not cope with well. The first is that of latency when interacting with data, the second is that of data affinity.

Firstly lets categorise the types of applications that bring these issues to the Grid as being ‘stateful’ applications and those that are traditionally deployed on the Grid as being ‘stateless’. Stateful applications bring additional challenges that compute Grid were not designed to cope with, namely:

- Data Contention: When fetching data for Grid nodes and when saving data.
- Data Affinity: Ensuring the correct pieces of data are delivered to the correct node on the Grid so that a task has the data it requires to execute
- Latency: When dealing with stateful applications on the Grid we hit Amdahl’s law , which essentially presents a theoretical upper limit on how performance can be increased when having to save state.

You can think of Enterprise DataGrid as a specialized type of Grid that can deal with stateful as well as stateless applications that are deployed onto it.

Data caching or distributed data caching is often used interchangeably with Enterprise Data Grid (EDG). Whereas the two have similarities to my mind they are different. They have similarities because vendors market products that seem to do the same thing i.e. providing an in-memory distributed data cache. However an EDG has to plug seamlessly into a Grid, be able to provide data affinity, be able to be used by the Grid Scheduler, be able to handle multi-tenancy, be able to be managed by the Grid management tools. There are facets that are the same but facets that mark out an EDG as something more than just a distributed cache.

How does this apply to the mobile world ?  Well, this type of technology is often the backbone to the services that are provided by Telco's.  A good example and description of this can be found in this post by Julian Browne. The Technology also underpins the delivery of mobile services in other industries outside of just finance and Telco.  A good example of this can be found in this case study .

Read the original blog entry...

IoT & Smart Cities Stories
The platform combines the strengths of Singtel's extensive, intelligent network capabilities with Microsoft's cloud expertise to create a unique solution that sets new standards for IoT applications," said Mr Diomedes Kastanis, Head of IoT at Singtel. "Our solution provides speed, transparency and flexibility, paving the way for a more pervasive use of IoT to accelerate enterprises' digitalisation efforts. AI-powered intelligent connectivity over Microsoft Azure will be the fastest connected pat...
There are many examples of disruption in consumer space – Uber disrupting the cab industry, Airbnb disrupting the hospitality industry and so on; but have you wondered who is disrupting support and operations? AISERA helps make businesses and customers successful by offering consumer-like user experience for support and operations. We have built the world’s first AI-driven IT / HR / Cloud / Customer Support and Operations solution.
Codete accelerates their clients growth through technological expertise and experience. Codite team works with organizations to meet the challenges that digitalization presents. Their clients include digital start-ups as well as established enterprises in the IT industry. To stay competitive in a highly innovative IT industry, strong R&D departments and bold spin-off initiatives is a must. Codete Data Science and Software Architects teams help corporate clients to stay up to date with the mod...
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...
Druva is the global leader in Cloud Data Protection and Management, delivering the industry's first data management-as-a-service solution that aggregates data from endpoints, servers and cloud applications and leverages the public cloud to offer a single pane of glass to enable data protection, governance and intelligence-dramatically increasing the availability and visibility of business critical information, while reducing the risk, cost and complexity of managing and protecting it. Druva's...
BMC has unmatched experience in IT management, supporting 92 of the Forbes Global 100, and earning recognition as an ITSM Gartner Magic Quadrant Leader for five years running. Our solutions offer speed, agility, and efficiency to tackle business challenges in the areas of service management, automation, operations, and the mainframe.
The Jevons Paradox suggests that when technological advances increase efficiency of a resource, it results in an overall increase in consumption. Writing on the increased use of coal as a result of technological improvements, 19th-century economist William Stanley Jevons found that these improvements led to the development of new ways to utilize coal. In his session at 19th Cloud Expo, Mark Thiele, Chief Strategy Officer for Apcera, compared the Jevons Paradox to modern-day enterprise IT, examin...
With 10 simultaneous tracks, keynotes, general sessions and targeted breakout classes, @CloudEXPO and DXWorldEXPO are two of the most important technology events of the year. Since its launch over eight years ago, @CloudEXPO and DXWorldEXPO have presented a rock star faculty as well as showcased hundreds of sponsors and exhibitors! In this blog post, we provide 7 tips on how, as part of our world-class faculty, you can deliver one of the most popular sessions at our events. But before reading...
DSR is a supplier of project management, consultancy services and IT solutions that increase effectiveness of a company's operations in the production sector. The company combines in-depth knowledge of international companies with expert knowledge utilising IT tools that support manufacturing and distribution processes. DSR ensures optimization and integration of internal processes which is necessary for companies to grow rapidly. The rapid growth is possible thanks, to specialized services an...
At CloudEXPO Silicon Valley, June 24-26, 2019, Digital Transformation (DX) is a major focus with expanded DevOpsSUMMIT and FinTechEXPO programs within the DXWorldEXPO agenda. Successful transformation requires a laser focus on being data-driven and on using all the tools available that enable transformation if they plan to survive over the long term. A total of 88% of Fortune 500 companies from a generation ago are now out of business. Only 12% still survive. Similar percentages are found throug...