All across the Internet, the Border Gateway Protocol, or BGP, is used to direct network traffic from one site to another. Here’s a look at how BGP works…
As the Internet economy experiences a major resurgence, utilization is soaring and new applications are impacting every aspect of consumers’ lives, from dating to social networking to entertainment.
Summary: Google has changed its service level agreement for paid versions of Google Apps, its suite of online applications. The goal, says Google Enterprise product management director Matt Glotzbach, is to deliver service that’s as reliable as telephone dial tone.
Google hasn’t been as consistently reliable as that. Its Postini e-mail archiving service was down for about a day in October 2009. There was a 100-minute Gmail outage in September 2009 and a more significant one in May of that year, as well as several other such incidents. Downtime happens, to Google, Microsoft, and just about every other cloud service provider you can think of. But with millions of enterprise customers, Google aims to become more reliable.
Jeffrey’s Thoughts: The comparison to the telephone dial tone as a “utility” may not have been the best choice since the phone company doesn’t take telephones down for routine maintenance. That is a pretty big qualifier for an application hosting provider that wants to be taken seriously in the enterprise space.
Speaking of qualifiers – 32 times more reliable that the average email system is not quite correct since Google is not addressing the fact that internet access is required to access Gmail. If you factor in the users’ ISP downtime to Google’s downtime that should be considered the TOTAL downtime for Google – uptime for Google’s services are irrelevant if I can’t get to them.
Google should look at a “Total Downtime of Ownership”, like a total cost of ownership, and include all of the things outside of Google that have to be in sync for their hosted applications to work. It’s statistics like these which make me believe 87.423% of statistics are made up on the spot. Shame on you Google for misleading folks that if your servers work, but no one can get to them, then they should be happy with how the system operates. When I’m at my office, at least I can still access our mail server when the internet is down.
So let’s look at the other uptime numbers Google shares – 99.7% Postini uptime (a little over 2 hours of downtime per month) and 99.98% uptime (almost 9 minutes downtime per month) for Gmail is still a very far cry from 100%. While I appreciate Google talking about improving uptime and providing an SLA for enterprise services – i have to ask – isn’t that a requirement?
Summary: $500 billion — it’s a number so big you’d assume it’s a component of the national debt. It isn’t. Instead, it’s what Gartner analyst Andy Kyte calls the IT debt. “When budgets are tight, maintenance gets cut. After a decade of tight budgets, the scale of the maintenance backlog has created systemic risk, particularly for large organizations,” he says. The “debt” really has two major components: One is underfunding and even neglect of routine but important hardware replacement purchases and software upgrades. The other is the slow degradation of enterprise applications. Is $500 billion too high? Kyte estimates that a typical Fortune 2000 company would require upgrades costing more than $200 million each.
Jeffrey’s Thoughts: I laughed out loud when I read about IT managers assuming the “cloud” will save them money in the not too distant future. Amazon EC2 is *MORE* expensive than fully managed dedicated hosting with full OS and application support if you use it more than 10%/month. For Windows customers, it’s only 7%/month. Don’t believe me? I’m happy to send you the numbers.
Bill Snyder’s example of Sparks, Nevada losing 42% of their staff is actually pretty light based on our experience with the companies we’ve consulted for. The reduction of staff has been frequent with no new job postings for these companies. Everyone who works in IT now should *expect* their work load to increase and stay high. CFO’s are demanding (out of necessity or simply being opportunistic) for IT to do more with less and even after the economy improves, I would not expect this attitude to change. Why? Companies showing profitability and margin through solely cost cutting are not going to give that up when they don’t have to. The CFO’s focus is on increasing shareholder value and bottom line profitability. The sooner we IT peons accept this is the new normal, the sooner we’ll at least be honest with ourselves.
The $500 Billion seems ridiculously high and geared to grabbing headlines. One observation – with all of the staff reductions, how have those savings resulted in less storage requirements, more spare laptops and desktops, and lower IT support loads simply because there are not as many people around? Also, with more applications being outsourced (I.e. CRM, email, ERP, etc) how does that in any way require maintenance? Also, weren¹t we all elated to skip Vista? I wouldn’t call that deferred maintenance, just common sense.
Summary: Cloud computing has many benefits for end users. But can it make data center providers more profitable? Financial metrics show that the growing cloud computing operation at Rackspace Hosting is helping the company deliver more revenue from its data centers. Over the past year, Rackspace (RAX) has added 38,000 customers, nearly all of them in its fast-growing cloud hosting business.
For those of you who missed Jeffrey’s session from the recent Gartner ITxpo in Orlando, here is his presentation on video -thanks to Hitachi Data Systems. It gives some great insights on network and datacenter convergence as well as how cloud solutions play a role in the IT landscape.
As always, if you’d like a copy of Jeffrey’s presentation slide, or for more information, feel free to contact us. Also, feel free to check out the official site of the conference, as well as Hitachi’s take on the event.
Summary: The federal government Tuesday afternoon released draft plans for a program to ensure cloud services meet federal cybersecurity guidelines, which should help shore up lingering government concerns about cloud security and accelerate adoption of the technology. The government expects the program, the Federal Risk and Authorization Management Program (FedRAMP), to be operational by the first quarter of 2011.
Our Thoughts: While these are just in draft form – these are security guidelines that should have been included in any “cloud” computing model from its inception. To not address the security needs in a computing business model means significant challenged to retroactively add them in. Given the multiple layers of abstraction between hardware and conventional shared infrastructure powering typical clouds – this promises to be a challenge.
Summary: My colleague Mark Bokwer and I are at a Virtualization, Cloud Computing, and Green IT conference in Washington DC this week. In one of the panels we hosted, an IT executive from a cabinet-level agency mentioned that the agency was qualifying Microsoft Hyper-V even though it already has an enterprise license in place with VMware. When asked why the agency was doing this he responded, “we are a Windows shop and have a great relationship with Microsoft. VMware has been great but we simply believe that the world is moving to heterogeneous virtualization platforms and we want to be ready for this.”
Our Thoughts: This merely mirrors what we see in the enterprise clients we have worked with. Vendor lock-in continues to be a concern with virtualization and given their variety of price points and licensing models, it’s smart to have exposure to the various virtualization platforms out there. In our experience the management tools for the platforms can vary widely, and the underlying technology can provide some insight into the most scalable platform for a particular application. Having experience with a given environment in several virtualization platforms can even help identify areas for further performance improvements.
Summary: Hitachi Data Systems has added remote office cacheing facilities to its multi-tenant Hitachi Content Platform (HCP) archive product, and allowed tenants to sub-let with individual archive policies for the sub-lettees. Version 4.0 increases the granularity of the multi-tenant feature added to HCP last year in v3.0. A user or tenant of the HCP can now have multiple namespaces with each namespace providing storage for a component of the tenant’s organization, such as order-processing, manufacturing, sales, etc.
Our Thoughts: Peak Web has been taking advantage of the HCP archive product and providing custom archiving and storage solutions on top of it. The multi-tenant capabilities were a key selling point – not only allowing our clients to securely share the environment, but also allowing them to allocate individual resources and policies per internal guidelines. The addition of remote office cacheing also beings another dimension to the HCP since it can serve as a central or distributed repository for offsite data storage. We are looking forward to seeing this in action this year!
Summary: The data center market experienced a boom in the first decade of the 21st century, but this boom came with a price. We are now in the “age of regulation,” and government compliance is becoming increasingly crucial for data centers. Public companies, by nature, are required to be Sarbanes-Oxley (SOX) compliant, and both public and private firms, are required to be HIPAA compliant. As such, data center operators have budgeted for huge costs and workloads for SAS 70 Audits. SAS 70 audits satisfy the SOX 404 reporting requirement, which requires an annual report on the effectiveness of the organization’s internal controls.
Our Thoughts: Compliance requirements can be difficult to navigate when implementing, but are a helpful point of view into the internal health of an organization. Since datacenters typically house critical information (usually for a variety of tenants) – having a SAS 70 Type II audit can provide some peace of mind. The Type II report is most helpful since it also includes a third-party auditors opinion on the level of controls in place and their effectiveness. While no compliance report is a replacement for due diligence on your next datacenter move – it’s an essential requirement to deal with the increase in regulation.
Peak Web Consulting came in during an especially critical growth phase in our company. Their network engineering and support helped grow our enterprise network by over 300% and increased overall quality of service to our customers, partners and audience. Their combined knowledge and skillset has helped save us time after time again, from both a proactive planning standpoint as well as during those unexpected emergencies. I would highly recommend Peak Web for any complex network engineering project.
Peak Web Consulting has provided our company advanced network consulting without having to keep a full time senior engineer on staff. They have trained several of our employees on basic networking knowledge so everyday problems can be dealt with immediately. Peak Web Consulting has provided responsive and reliable service at a very reasonable price. We continue to be satisfied with their performance and can whole heartily recommend them to others.
John Wen Director of Systems Engineering JDate.com .
