Categories
Comment

Google PR agency to feedback on the long-term availability of the Google Compute Engine

After I have called the long-term availability of Google’s Compute Engine (GCE) into question, Google’s PR agency has contacted me to understand the motivations for the article. In the article I have reacted to a GigaOM interview of Googles Cloud Platform manager Greg DeMichillie who wouldn’t guarantee the long-term availability of GCE.

Background

Google Cloud Platform manager Greg DeMichillie responded in a GigaOM interview to a question on the long-term availability of Google cloud services and answered unexpected and not within the meaning of the customer.

„DeMichillie wouldn’t guarantee services like Compute Engine will be around for the long haul, but he did try to reassure developers by explaining that Google’s cloud services are really just externalized versions of what it uses internally. ”There’s no scenario in which Google suddenly decides, ‘Gee, I don’t think we need to think about storage anymore or computing anymore.“

Although DeMichillie to qualify in the end that Google wouldn’t shut down their cloud services in a kamikaze operation. However, it’s an odd statement on a service which is relatively as of late on the market.

Feedback of Google’s PR agency

Beforehand I want to clarify that this was no call to influence me, but to understand how it came to the article. Google’s PR agency said that DeMichillie was apparently misunderstood and my article only highlights this negative statement and neglects the positive themes. Google today severely invests in infrastructure resources and there are no signs and reasons that the Google Compute Engine will be closed.

My statement

It was and is never about to cast a shadow on Google (or any other vendor). This is what I told the PR agency. But at the end of the day the user needs to be advised and equally be protected. Moreover I am an analyst and advisor and counsel companies who rely on my judgment. For this reason I need to react on those statements and include it in my decision matrix, especially when it directly comes from an employee of a vendor. What should I do when I recommend the use of the GCE since the technical things and requirements fits, but Google subsequently announced to close the service? For this reason I react extremely sensitive on such topics. In addition, Google had not cover oneself in glory the recent months and years when it comes to maintain its service portfolio for the long-term. The end of Google Reader caused more negative reactions by the users than the current NSA scandal. Notabene, this is a free service for consumers. With the Google Compute Engine we are talking about a service which mainly address companies. For companies it’s about a lot of money to spend to bring the workloads to the GCE. If the service is suddenly closed this generates, depending on the respective company, a non incalculable economic damage to migrate the data and applications. This Google should consider when making decisions. Even if this was just one statement as part of an interview. This does not create trust in the cloud portfolio and fits well with the experience of the recent past when Google cleaned it’s room.

Categories
Comment

Google Compute Engine seems to be no solution for the long haul

In an interview with GigaOM, Google‘s Cloud Platform manager Greg DeMichillie made an odd statement on the future of the Google Compute Engine which again have to lead to a discussion on the future-proofness of Google’s cloud service portfolio and if it makes sense to depend on the non core business areas of the search engine provider.

Google is to agile for its customers

After Google announced to close the Google Reader, I already asked the question how future-proof the Google cloud portfolio is. In particular, because of the background, that Google starts to monetize more and more services, those due to the revenue get a new KPI and thus are threaten a closure. Google Cloud Platform manager Greg DeMichillie exactly meets this question in a GigaOM interview and answered unexpected and not within the meaning of the customer.

„DeMichillie wouldn’t guarantee services like Compute Engine will be around for the long haul, but he did try to reassure developers by explaining that Google’s cloud services are really just externalized versions of what it uses internally. ”There’s no scenario in which Google suddenly decides, ‘Gee, I don’t think we need to think about storage anymore or computing anymore.“

Although DeMichillie to qualify in the end that Google wouldn’t shut down their cloud services in a kamikaze operation. However, it’s an odd statement on a service which is relatively as of late on the market.

These are things customers should better not hear

The crucial question is why a potential customer should decide for the Google Compute Engine for the long haul? Due to this statement one have to advise against the use of the Google Compute Engine and instead set on a cloud computing provider who has its actually core business in infrastructure-as-a-service and not be indulgent to sell its overcapacities and instead operate a serious cloud computing business.

I don’t want to speak of the devil and the devil shows up! But news like the sudden death of Nirvanix – an enterprise cloud storage service – to make massive waves and outface the users. This also Google should carefully understand if it wants to become a serious provider of cloud computing resources.

Categories
Analysis

Google expands the Compute Engine with load balancer functionality. But where is the innovation machine?

In a blog post, Google has announced further renovations to its cloud platform. In addition to the expansion and improvement of the Google Cloud Datastore and the announcement of the Google App Engine 1.8.3, the infrastructure-as-a-service (IaaS) offer Google Compute Engine has received a load balancing service.

News from the Google Cloud Datastore

To further increase the productivity of developers, Google has expanded its Cloud Datastore with several functionalities. These include the Google Query Language (GQL). The GQL is a SQL-like language and is specifically aimed at data-intensive applications to query Entities and Keys from the Cloud Datastore. Furthermore, more statistics can get from the underlying data by the query of metadata. Google sees this as an advantage, for example, to develop an own internal administrative console to do own analysis or debug an application. In addition, improvements to the local SDK were made. These include enhancements to the command line tool and support for developers using Windows. After the first version of the Cloud Datastore has support Java, Python, and Node.js, now Ruby has been added.

Google App Engine 1.8.3

The renewal of the Google App Engine version 1.8.3 mainly directed to the PHP runtime environment. Furthermore, the integration with the Google Cloud Storage, which according to Google is gaining popularity, was reinforced. From now on functions such as opendir() or writedir() can be used to directly access bucket in Cloud Storage. Further stat()-ing information can be queried via is_readable() and is_file(). Metadata can also be write to Cloud Storage files now. Moreover performance improvements through memcache-backed optimistic read caching were made, to improve the performance of applications that need to read frequently from the same Cloud Storage file.

Load Balancing with the Compute Engine

The Google Compute Engine is expanded with a Layer 3 load balancing functionality which allows to develop scalable and fault-tolerant web applications on the Google IaaS. With the load balancing function, the incoming TCP/UDP traffic can be distributed across different Compute Engine machines (virtual machines) within the same region. Moreover, this ensures that no defective virtual machines are used to respond to HTTP requests and load peaks are offset. The configuration of the load balancer is carried out either via command line or via the REST API. The Load Balancer function can be used free of charge until the end of 2013, then the service will be charged.

Google’s progress is too slow

Two important innovations stick out in the news about the Google Cloud Platform. On the one hand, the new load balancing function of the Compute Engine, on the other hand the strengthen integration of the App Engine with Google Cloud Storage.

The load balancing extension is admittedly much too late. After all, this is one of the essential functions of a cloud infrastructure offering, to ensure to develop scalable and highly available systems and which all other existing providers on the market already have in their portfolio.

The extended integration of the Cloud Storage with the Google App Engine is important to give developers more S3-like features out from the PaaS and create more opportunities in terms of access and to provide the processing of data in a central and scalable storage location.

Bottom line, it can be stated that Google continues to expand its IaaS portfolio steadily. What stands out here, is the very slow speed. From the innovation engine Google, which we know from many other areas of the company, is nothing to see. Google gives the impression that it has to offer IaaS to not lose too many developers on the Amazon Web Services and other providers, but the Compute Engine is not highly prioritized. As a result, the offer has treated very little attention. The very late extension to the load balancing is a good example. Otherwise you might expect more ambition and speed to invest in the expansion of the offer, from a company like Google. Finally, except for Amazon, no other company has more experience in the field of highly scalable infrastructures. This should be relatively quickly transformed into a public offering, especially like Google advertises that customers can use the same infrastructure on which also all Google services are running. Apart from the App Engine, which is classified in the PaaS market, the Google Compute Engine must continue to hire back. This will be the case for a long time, if the expansion speed is not accelerated, and other essential services are rolled out.