TOKYO -- Hitachi will supply IBM-made mainframe computers loaded with its own operating systems starting in fiscal 2018, beating a retreat from hardware development in the age of cloud servers.

The arrangement was announced Tuesday. Hitachi will continue developing operating systems for mainframe machines.New products will offer improved compatibility with Hitachi's "internet of things" platform Lumada.

Mainframes have been widely used across Japan's public and private sectors for in-house computer systems since they emerged in the 1950s. Japan was a particularly big market, with Hitachi, IBM, NEC, Fujitsu and others all competing for a piece of the pie at one point.

But the tide turned in the 1990s, when computer servers loaded with Windows and Linux operating systems became widespread. Japan's mainframe shipments topped 1 trillion yen ($8.94 billion) in the mid-1990s, but slid to below 45 billion yen in fiscal 2015, according to the Japan Electronics and Information Technology Industries Association.

That changing market climate prompted Hitachi's eventual exit from hardware. At the same time, the company recognizes persistent demand among companies that value stable system operation and security. There is also promise for use in the internet of things, a business Hitachi sees as a growth field.

The turning fortunes of IBM are symbolic of the technological shift. The company's sales fell for a 20th straight quarter in the January-March period of this year. The company was hit by the advent of servers and then by the spread of cloud servers.This is in stark contrast to Amazon.com's cloud business, which is the world's largest. The business logged $890 million in operating profit in the January-March period, accounting for 90% of the company's profit.

In Japan, where cloud computing has not become as widespread, IT companies have been logging relatively solid earnings. Yet, Fujitsu, NEC and Hitachi all saw revenue in the IT segment decline in the year ended in March. Making it imperative for them to respond to the growth of cloud computing.

(Nikkei)

Excerpt from:
Hitachi, IBM to collaborate in mainframes in the cloud era - Nikkei Asian Review

Thank you

Your message has been sent.

There was an error emailing this page.

By Paul Krill

Editor at Large, InfoWorld | May 22, 2017

Firebase, Google Clouds back end and SDK for mobile and web application development, is being enhanced with serverless compute capabilities. Google Cloud Functions for Firebase, now available in a beta release, allows developers to run back-end JavaScript code that responds to events triggered by Firebase features and HTTPS requests.

Developers upload their code to Google's cloud, and the functions are run in a managed Node.js environment. There is no need for users to manage or scale their own servers.[Cloud Functions] enables true server-less development, Google's Ben Galbraith said. LikeAWS Lambda and Microsoft's Azure Functions, Cloud Functions allows users to deploy and run code without provisioning servers. Developers code to cloud APIs, and the cloud takes care of managing and scaling the functions.

Acquired by Google in 2014, Firebase features a cross-platform SDK with capabilities for cloud data storage and synchronization across devices. It also provides app usage analytics and tools for serving in-app advertising and sending targeted notifications to users.

Google has also just released a beta version of Firebase Performance Monitoring. The service provides insight into the performance of iOS and Android mobile apps by monitoring startup times, network response times, and other aspects of app performance. The data can be analyzed in the Firebase Console.

Google also has begun open-sourcing Firebase SDKs, describing it as the first step toward open-sourcing client libraries. Were starting by open sourcing several products in our iOS, JavaScript, Java, Node.js, and Python SDKs. We'll be looking at open sourcing our Android SDK as well, Salman Qadri, Firebase product manager, said. Admin SDKs to access Firebase on privileged environments are now open source as well, including the recently launched Python SDK.

Paul Krill is an editor at large at InfoWorld, whose coverage focuses on application development.

Sponsored Links

Excerpt from:
Google's Firebase taps serverless Cloud Functions - InfoWorld

Only a week after Nvidias new AI-focused Volta GPU architecture was announced, Google aims to steal some of its thunder with its new, second-generation, Tensor Processing Unit (TPU) that it calls a Cloud TPU. While its first generation chip was only suitable for inferencing, and therefore didnt pose much of a threat to Nvidias dominance in machine learning, the new version is equally at home with both the training and running of AI systems.

At 180 teraflops, Googles Cloud TPU packs more punch, at least by that one measure, than the Volta-powered Tesla V100 at 120 teraflops (trillion floating point operations per second). However, until both chips are available, it wont be possible to get a sense of a real world comparison. Much like Nvidia has built servers out of multiple V100s, Google has also constructed TPU Pods that combine multiple TPUs to achieve 11.5 petaflops (11,500 teraflops) of performance.

For Google, this performance is already paying off. As one example, a Google model that required an entire day to train on a cluster of 32 high-end GPUs (probably Pascal), can be trained in an afternoon on one-eighth of a TPU Pod (a full pod is 64TPUs, so that means on 8TPUs). Of course, standard GPUs can be used for all sorts of other things, while the Google TPUs are limited to the training and running of models written using Googles tools.

Google is making its Cloud TPUs available as part of its Google Compute offering, and says that they will be priced similar to GPUs. That isnt enough information to say how they will compare in cost to renting time on an Nvidia V100, but Id expect it to be very competitive. One drawback, though, is that the Google TPUs currently only support TensorFlow and Googles tools. As powerful as they are, many developers will not want to get locked into Googles machine learning framework.

While Google is making its Cloud TPU available as part of its Google Compute cloud, it hasnt said anything about making it available outside Googles own server farms. So it isnt competing with on-premise GPUs, and certainly wont be available on competitive clouds from Microsoft and Amazon. In fact, it is likely to deepen their partnerships with Nvidia.

The other company that should probably be worried is Intel. It has been woefully behind in GPUs, which means it hasnt made much of a dent in the rapidly growing market for GPGPU (General Purpose computing on GPUs), of which machine learning is a huge part. This is just one more way that chip dollars that could have gone to Intel, wont.

Big picture, more machine learning applications will be moving to the cloud. In some cases if you can tolerate being pre-empted its already less expensive to rent GPU clusters in the cloud than it is to power them locally. That equation is only going to get more lopsided with chips like the Volta and the new Google TPU being added to cloud servers. Google knows that key to increasing its share of that market is having more leading edge software running on its chips, so it is making 1,000 Cloud TPUs available for free to researchers willing to share the results of their work.

The rest is here:
Google Targets Nvidia With Learning-Capable Cloud TPU - ExtremeTech

Comment Cisco has missed out on a blade to rack server shift, sales growth has turned negative, it doesn't sell to cloud providers and it has a small market share. Should it invest to grow or get out of servers altogether?

Cisco's third fiscal 2017 quarter results were disappointing, with a 1 per cent decline in revenue year-on-year to $11.9bn. The data centre segment, meaning UCS servers mostly, made $767m in revenue and was down 5 per cent. It constitutes just 6 per cent of Cisco's overall revenues.

In the previous quarter data centre revenues were $790m, down 4 per cent year-on-year, and in the quarter before that they were $834 million, down 3 per cent year-on-year. There is a pattern of decline here.

Stifel analyst and MD Aaron Rakers has charted this, showing Cisco data centre revenues and the year-on-year percentage change:

The chart shows actual numbers plus estimates looking at nine quarters.

UCS servers blazed a bright trail in the sky when they first arrived. What is going on?

Overall server sales are down, according to both Gartner and IDC. Dell and HPE lead the market, followed by IBM, Lenovo and Huawei.

IDC gave Cisco a 6.3 per cent market share in 2016's fourth quarter, with HPE having a 23.6 per cent share, Dell 17.6 per cent, IBM 12.3 per cent and Lenovo 6.5 per cent. Original design manufacturer (ODM) suppliers accounted for 7.9 per cent. Why is Cisco lagging?

Rakers charted quarterly server sales by architecture over the past few years:

Rack-optimised server sales are the big winners, with blade server sales second, a long way behind, and growth stopping. Density-optimised server sales are flattish, towers are in decline and large systems are the smallest category, although growing slightly.

Rakers next plotted Cisco's UCS server sales in the blade and rack segments, showing both revenues and revenue share percentages:

Most of Cisco's UCS revenues come from blade server sales, the declining second-placed architecture, and not rack servers, the main and growing segment. The conclusion is inescapable; Cisco has misread the server market badly, with revenue growth slowing drastically and then stopping from its first fiscal 2015 quarter, two and a half years ago.

Rakers said: "Cisco continues to face a misaligned portfolio for the mix from blade to rack servers i.e. Cisco has ~30 per cent revenue share in blades; sub-4 per cent share in rack servers."

Cisco sells its servers to enterprises and not the hyperscalers or cloud service providers, ODMs such as Supermicro and Chinese server suppliers such as Inspur.

Cisco has been pushing its HyperFlex hyperconverged infrastructure appliance (HCIA), using OEM'd Springpath software. In March Cisco said it had gained 1,100 HyperFlex customers after nine months of sales. Nutanix has around 5,400 and we expect Dell EMC to be in that kind of area soon.

In its third-quarter results announcement Cisco did not update the 1,100 customer number. A Stifel survey of Cisco's VARs/resellers found 16 per cent thought HyperFlex was best positioned in the HCIA market while 40 per cent thought Nutanix was the leader. Some 66 per cent had sold HyperFlex systems into existing Cisco accounts, not new customers.

Rakers said that some 20 per cent of server revenues come from sales into the public cloud, and Cisco does not sell there, with ODMs and white box servers having around a 40 per cent share.

To sum up, Cisco's servers account for 6 percent of its overall revenues, and these revenues have been declining for four quarters in a row. It has a 6.3 per cent share of the overall market, but a less than 5 per cent share in the biggest and growing rack server section. It's progress in the HCIA market was off to a good start but it lags a long way behind market leaders Nutanix and Dell. HPE, by buying SimpliVity, is becoming a stronger competitor.

Finally it is not a supplier to the public cloud server marker.

It seems to us that, to make progress with servers, Cisco needs to get into rack servers in a big way. But there is a more fundamental question; what is its goal here? Does it want to be a leading server supplier, up with Dell and HPE? Or is it content to have a sub-10 per cent share of the market, selling into its installed base and under continual attack from Dell, HPE and the various Chinese and ODM suppliers?

If it wants to get up with the leaders then it has to spend a lot of money on engineering development and so forth. That will be a hard call when overall revenues are declining, servers are just 6 per cent of its business and it's laying people off.

Perhaps Cisco should step back, take a deep breath, and decide to exit the server market, selling its UCS business to Lenovo, say. Perhaps on the other hand it could try something radical, like buying Supermicro.

That doesn't feel right with our view of Cisco moving into servers as an adjacent market to its core networking market. It then moved into storage as an adjacent market to servers, and failed. We think Cisco sees its server market prospects as being limited, and can't see it making the investments needed to become a top four or five server supplier.

Looking ahead we reckon there'll likely be product line tweaking, statements of renewed commitment and determination, but little actual change in its situation. Servers are too large a part of Cisco's revenues to throw away, too small a part to be worth investing heavily in, and not in a dire enough situation to need a radical fix. They're stuck in limbo and look likely to stay there.

See more here:
Cisco's servers are stuck in limbo, look likely to stay there - The Register

Nvidia: This Could Work Out Great, Says Bernstein - Barron's Barron's Bernstein analyst Stacy Rasgon started coverage of GPU chip maker Nvidia with the equivalent of a Buy rating, and a $165 price target, arguing its market for ... Nvidia makes the case for GPU accelerators | ZDNetZDNet all 2 news articles »

Read the rest here:
Nvidia: This Could Work Out Great, Says Bernstein - Barron's - Barron's