Technology Update

Microsoft launches Server 2008 SMB options

noreply@blogger.com (Seth) — Wed, 12 Nov 2008 20:48:00 +0000

Microsoft made great inroads in the SMB space with their Small Business Server solution. It provides the server and application licenses at a much lower price than buying the full products outright. There are some limitations to it – most notably the 75 user limit – but in many cases it is the best solution for an organization. Today Microsoft is extending those benefits to the Server 2008/Exchange 2007 platform with the launch of Windows Essential Business Server 2008 (“EBS”) and Windows Small Business Server 2008 (“SBS”).

SBS remains as a single server solution (two servers in the premium edition), but now with the most current versions of the server and application suites. It also remains capped at 75 users. EBS is a new product offering from Microsoft for slightly larger organizations. The basic edition includes multiple server licenses (up to three) and allows for up to 300 users to be licensed.

The pricing on the licenses is still better than buying the individual server licenses and CALs outright, but the limitations are hard to overlook. For an organization with very simple and basic needs these products are fine. If, however, an organization needs any flexibility or ability to grow the SBS/EBS options are rather limiting. Plus, I dislike the fact that Microsoft assumes that an SBS/EBS customer would necessarily use all the components and how things actually don’t work right if you don’t install them all.

This release is great news for SBS customers who want to take advantage of the benefits Exchange 2007 provides. Unfortunately Microsoft’s documentation for migrating to the new platform is still pending, so it is pretty much a crap-shoot at this point. Good luck with that.

Also on the good news front is that SBS/EBS can be purchased with Server 2003 R2 instead of Server 2008 until December 31, 2009 for folks who are a bit gun-shy about running the new server OS. Similarly SQL 2005 is still an option rather than SQL 2008 for the premium versions.

AT&T opens up WiFi network to smartphones

noreply@blogger.com (Seth) — Sat, 01 Nov 2008 20:27:00 +0000

AT&T has decided to open up their WiFi network for free to anyone who has a smartphone on a data contract with the carrier. The offering was originally for iPhone users but it has also been announced for Blackberry Bold users and then for all smartphone contracts.

Why is AT&T doing this? I guess it depends on who you ask, but most theories seem to be focused on this being an offering to make up for poor 3G performance on their network and many upset iPhone users. With the now November 4 release date of the Bold having been delayed more than once because of concerns about the 3G network performance. Since the AT&T WiFi coverage includes all Starbucks locations they’re basically offering a huge hotspot solution to the bandwidth problem. Hopefully it will assuage their customers. If it doesn’t their approach of having cool phones available first may not be enough to keep their business growing.

Sprint and Cognet in a peering battle

noreply@blogger.com (Seth) — Sat, 01 Nov 2008 00:30:00 +0000

I spent a couple hours today troubleshooting a problem with connectivity from a client to a collection of web sites. The cause? Apparently Cogent went and upset another ISP with their cheap bandwidth and decent support services. This time around it is Sprint that is upset, and their solution was to black hole all traffic destined to the Cogent network.

If this sounds familiar that’s because it is. Cogent has had similar problems in the past with other ISPs and has resolved them eventually, but the outages are an incredible pain to the affected customers. A law firm unable to file documents with a court based on their ISP choice could be somewhere between malpractice and disastrous. And there is very little that a customer can do about it one they find themselves cut off.

Of course, planning ahead provides a chance of success. In this particular case having redundant internet connections from different carriers and BGP configured for IP connectivity meant that forcing traffic to the “other” line was sufficient to get around the problem, with virtually zero impact on the end users. This may not be a “typical” business continuity scenario since everything in the world appears to be operating normally, but for this client having such a configuration means overcoming a significant barrier to doing business based on circumstances outside of their control. If that isn’t business continuity in a nut shell I don’t know what is.

Not sure if your business continuity plan is up to snuff? Let Proton Associates talk it over with you and see how it looks.

Windows 7 shows up at PDC

noreply@blogger.com (Seth) — Tue, 28 Oct 2008 21:02:00 +0000

We knew it was coming, and Microsoft delivered, showing off a “pre-beta” version of Windows 7 today at the PDC conference. And how does it look? Well, a lot like Vista, it seems. In fact, while I see a lot of press about “not making the same mistakes” and “simplifying things” for the users, there doesn’t seem to be much solid information about how they are going to do that. And since the basis of the new OS is shared with the core code of Vista, chances of major changes are pretty slim.

"The general perception is that Vista is a damaged brand, so it behooves Microsoft to move on to sell something new even if it's not a quantum leap in terms of technology," said Toan Tran, analyst at Morningstar.

In other words, if Microsoft can figure out how to release something with a new brand, even if it doesn’t really change much from Vista, Microsoft should do well by it. I have to disagree with this analyst. Customers want an OS that actually works, that is easily managed and easily deployed. Vista didn’t hit those targets, and unless things actually change under the covers Windows 7 will be just as disappointing as Vista, even if the hardware drivers all work just fine.

A “special” Microsoft Security alert

noreply@blogger.com (Seth) — Fri, 24 Oct 2008 01:50:00 +0000

So I’m sitting at home this evening, watching the World Series and I get an email from Microsoft. I get a lot of those so I rarely pay attention to them, but this one seems different. For starters, it says “critical” in the subject line, so I know it must be important. It turns out that they’ve released an out-of-cycle security patch that apparently affects every version of Windows.

What are the effects of the bug? The standard “Remote Code Execution” flaw. This one is in the Server service, which has been a part of Windows since the NT days. And the bug allows for unauthenticated user (except on Vista/Server 2008, where the user has to be authenticated) to compromise the system they are attacking completely. This is the the type of hole that major worms are based on. And even though the bug almost certainly affects systems dating back to NT4 and probably NT3.51, Microsoft only has patches for Windows 2000 SP4 and later versions because the older stuff is no longer supported. Awesome.

There’s a patch out, but no cool name for any exploits yet. So we don’t have a Slammer or a Melissa yet. But the hole is there. Go ahead and patch up all your computers. Good times.

Virtual Machine Manager 2008 released

noreply@blogger.com (Seth) — Wed, 22 Oct 2008 02:24:00 +0000

Microsoft has released the most significant piece of their move into the enterprise virtualization space today – Virtual Machine Manager 2008.

Sure, there is Hyper-V, and it is a decent hypervisor, but the hypervisor itself doesn’t really matter all that much. So when looking at a virtualization environment the management and integration tools are key. And Virtual Machine Manager is a huge leap forward on that front. Designed to integrate both into the System Center toolkit and with the three major commercial virtualization platforms (VMware, Citrix/Xen & Hyper-V), SCVMM08 is basically the cat’s pajamas of virtualization management. Get a copy of this tool if you can; it is definitely worth looking at if virtualization is part of your environment.

And for those of you keeping score at home, Citrix/Xen was busy earlier in the month, with the release of a beta version of Kensho, a tool to facilitate the move of virtual guests between hosts running different hypervisor platforms. Some additional comments about Kensho can be found in a post I had a couple months back.

Computing in the cloud

noreply@blogger.com (Seth) — Fri, 10 Oct 2008 16:23:00 +0000

“Cloud Computing” has been in the news a lot lately, almost as much as “virtualization.” And, just like virtualization, what it is and why it matters seems to vary wildly depending on who is doing the talking. For some, such as Oracle’s Larry Ellison, cloud computing is “complete gibberish,” and I cannot really argue with him based on some of the places I’ve seen the term used. Still, there are some companies that are making progress in the field and some of it actually looks pretty interesting.

The basic concept behind cloud computing is that all of the processing, storage and data transfer is billed as a commodity. Instead of owning a server on a shelf bits get processed and the bill is based on those cycles on the server. Even better, a cloud environment that is built out appropriately can allow for localized access to data around the world, effectively improving performance by replicating the data and application(s) in the background and serving them up from the closest point to the customer.

One of the biggest issues with the cloud, however, is that the existing solutions are proprietary and generally do not work with commodity software. In other words, applications have to be built to the specifications of the cloud company. That locks a customer in to the platform, which is a bad situation to be in. Even with Microsoft’s expected move into the space at the end of the month, the platform still appears to be proprietary. Amazon is apparently adding Windows-based servers to their EC2 platform, but it doesn’t appear that such a move is really for line-of-business applications as much as it is for web hosting and data warehousing solutions. I’ve been working with one company on a similar solution with a pretty wide open server solution that is completely transparent, allowing any server OS or application set to be run. But they don’t have the global coverage or background data migration working yet, so that’s not perfect either. That being said, it does make for a pretty impressive platform on which to run a Hosted Exchange environment.

Ultimately, the cloud computing initiative is just an extension of the recent drive to centralize all computing into a data center. No more paying for the individual components to build a network nor costs for the idle resources that aren’t being fully utilized. And if with the cloud solutions mirroring data between multiple locations business continuity comes as part of the package. That’s a big win, assuming such a solution really exists. I haven’t seen one fully baked yet, but it is getting much closer. And not having to build out a dedicated colo facility to get such benefits would be really nice.

WiMAX up and running in several cities

noreply@blogger.com (Seth) — Tue, 07 Oct 2008 02:37:00 +0000

WiMAX promises bandwidth and coverage capabilities that should be able to provide WiFi-like network coverage across cities rather than just hotspots. The technology has long been touted as a solution to wireless technology that is more reliable and better suited to fixed an mobile solutions than cellular or WiFi options. Unfortunately, no one has really been able to tell if that was the case because there have been no real deployments to speak of.

That changed last week when Sprint launched their Xohm service in Baltimore. At the time they also announced the next seven cities to receive service (Chicago, Washington, DC, Dallas, Fort Worth, Boston, Providence and Philadelphia) but without any particular timeline. It seems that several of these cities are actually online and running, albeit in a test mode right now. So if you have access from a service plan in Baltimore you can actually travel in several cities and use the service. No guarantees and it might not be 100% reliable, but it is good to see that the hardware is mostly in place and that the systems are up and running.

In addition to being useful for mobile users, WiMAX also provides an alternative to the traditional copper or fiber last mile options, so it can be considered as an alternate Internet connectivity option for an office. Latency might be an issue, but some testing should help figure that out. Still, not a bad DR/BC option if you’re in one of those cities.

Unified Communications might actually be happening

noreply@blogger.com (Seth) — Sat, 04 Oct 2008 23:47:00 +0000

I love the concept of Unified Communications (UC). The idea of a single inbox, a single phone number, integrated presence, IM, fax and just about everything else is pretty amazing. The problem is that there are a bunch of different standards and getting them all integrated has not always been easy. And with every company that touches anything remotely communications related piling on the UC bandwagon, the opportunity for failed integrations and other problems increases, which is not a good thing.

Cisco is one of the main culprits in the confusion in this space. They’ve assembled a pretty decent array of UC products now, from telephony to teleconferencing to instant messaging. And they continue to augment their portfolio with things like the purchase of Jabber, an IM player. But they have routinely used alternate protocols (SCCP instead of SIP in their telephony systems, for example) to keep their systems relatively closed off. Unfortunately this ignores one of the main benefits of UC, which is full integration and interoperability.

Fortunately, there has been some progress on the interoperability front from RIM and Microsoft. There is a client available to integrate the RIM devices (assuming a relatively current version of the server and client software) and Office Communications Server 2007. I’m a huge fan of Microsoft’s OCS 2007. It is the best implementation of a UC platform I’ve seen thus far, with hooks in to desktops, phones and mobile devices. And the ability to extend that to BlackBerry devices now is a great option. More than just a new IM client, the OCS client on the BlackBerry adds status information and voice integration to the OCS system.

Unified Communications is still viewed as a growth market more than an established one, but the good news is that the end game does seem to be integration and common platforms (even Cisco seems to have come around to SIP) rather than islands of communication.

Windows 7 sees daylight

noreply@blogger.com (Seth) — Thu, 02 Oct 2008 20:08:00 +0000

Sortof.

Later this month at the Microsoft Professional Developers’ Conference Microsoft is planning on distributing pre-Beta (Alpha??) bits of the Windows 7 platform. Does it really matter? Probably not. Considering that there was noise of a January 2010 release for Windows 7 it makes sense that there would be bits starting to show up now, though having only one year left to get through the Beta program and get the marketing wheels going seems a bit short, so I’m still betting on the delivery date slipping. But it is nice to see that they are making progress on the code.

Citrix confronts corporate computer challenges

noreply@blogger.com (Seth) — Fri, 26 Sep 2008 01:32:00 +0000

And they’re taking a rather interesting approach: Bring Your Own Computer. As part of a small pilot program (200 users) Citrix is giving folks a reasonable budget ($2100) and set of guidelines (Mac or Windows, 3-year full warranty and anti-virus subscription) and then letting them buy whatever they want. Claiming that they generally spend $25-2600 per person over the same time frame for a managed system, this move seems to be both about making the users happy and also about saving money.

By depending on the laptop vendors for supporting the system Citrix will off-load a pretty significant burden. And all of their business applications are already running in either a virtualized or Citrix environment, all with a pretty simple client deployment. But there is also the need to isolate an unknown/unmanaged system on the network, the potential for support calls to the main help desk anyway. But Citrix is hopeful.

The pundits are reasonably split among two camps. One says that this makes user happier and happy users are more productive. The other camp is not quite as convincing in their argument:

However, Sara Radicati, an analyst whose Radicati Group tracks business computing use, said she doesn't see what problem the Citrix program fixes, and she's unsure how useful it will be.

"We live in a complex world, so it is easier to manage and know where your data is and what is being done with company-sensitive information if you have a little more control," Radicati said.

Apparently she doesn’t realize that all that company-sensitive information is running in a thin client session. But there is the question of what problem this reall is addressing. Ultimately the success of the program will depend on Citrix extracting themselves from the management of the system, as that is going to drive the costs of the program up or down. And that still doesn’t take in to account the infrastructure changes that are probably prudent to accommodate the unmanaged workstations on the network. Sure, Citrix has some tools that help with that, so they should be able to get a good deal on the hardware costs, but there are still engineering costs that they have to deal with.

I’m on the side of the nay-sayers. I think that the management costs aren’t going to come down enough to really reduce the TCO numbers enough for this to be viable. And the lack of centralized management further reduces the benefit of this program. I guess we’ll see who’s right eventually.

Oh, and if you have a more traditional client-server application setup then this is a non-starter from day one, so don’t even think about it.

The biggest benefit of the G1 phone

noreply@blogger.com (Seth) — Thu, 25 Sep 2008 23:09:00 +0000

At least thus far, the most significant benefit of the device that has been realized is that TMobile has removed the bandwidth consumption cap from their data plan contracts. Considering that Verizon Wireless and Sprint both added caps on their “unlimited” plans, TMo choosing to remove it is a pretty interesting move.

Oh, and it looks like a decent device, too. Except for the part where TMobile’s 3G network is still pretty limited in size and runs on a different set of frequencies than any other 3G service in the rest of the world, meaning that global roaming with the device will be limited to EDGE speeds. And no Exchange support.

So it probably isn’t a very good business device, at least not until the Android software gets an update.

Update (10:36pm): I was wrong about the frequencies; it has 2100MHz support for Europe 3G, too. But still no Exchange support.

VMware’s grand plans

noreply@blogger.com (Seth) — Thu, 25 Sep 2008 02:30:00 +0000

VMworld was last week and I’m still trying to process most of what I’ve seen reported from it, but one thing is clear to me: VMware is still going full speed ahead on their product lines, though their focus has shifted away from the hypervisor. This makes sense, as the hypervisor is rapidly losing its value as a differentiator. They also are claiming that these efforts are the real deal, not just hype. They have stated that there are specific products to back up all these ideas and they are expected to be on the market by the end of 2009. That’s a long way off, but still a reasonable target to shoot for. So what are the big things coming? Here are a few…

Diverse location redundancy: The idea here is that you can run a production VMware environment in one location and mirror it to a second location and have an integrated fail-over solution without depending on third-party solutions. I’m skeptical based on potential issues with latency and client failover requirements, but it is certainly intriguing. Also, VMware isn’t the only group working on this front. So there is definitely some competition pushing for this to get done.

Network virtualization integration: Want a network switch integrated into your VMware platform? Want it to be more feature-rich than the basic vSwitch functionality that currently exists? You’re in luck, as this functionality is not only on the drawing board, but it actually appears to be somewhat matured. Cisco apparently has developed the necessary underpinnings for a software-only switch and they were discussing them during the week.

Virtual Desktop Infrastructure improvements: VDI is a decent concept for some applications, but the scalability on the hypervisor/server side isn’t really there yet for massive deployments and there are some other improvements to the concept that VMware is working on. This should be interesting, particularly if they get closer to attacking Citrix which it seems that they are planning on doing.

Bare-metal hypervisor for non-servers: I don’t know why anyone really cares too much about this, but VMware is making a fair amount of noise in the realm of a hypervisor designed for laptop/desktop hardware. Why this is particularly useful, I don’t know. Unless they are also going to add the ability to display the console of one of the running guests on the primary display (e.g. run VMware on the bare metal, a Vista session in that and view the Vista session on the screen rather than the VI3 console) I don’t really see the value to end users.

So, plenty of stuff going on and lots to look forward to in 2009 as these visions start to be realized and released.

Ike causes significant Internet outages

noreply@blogger.com (Seth) — Thu, 18 Sep 2008 01:26:00 +0000

In addition to all the other damage that Hurricane Ike caused last weekend as it barreled through Texas and middle America, Internet service suffered significantly. More than 400 different networks (ISPs, colleges, companies, etc.) were offline for at least one hour by Monday and some of them were still offline on Wednesday. This is the most significant outage since the 2003 blackout in the northeast, according to Renesys, a company that tracks network availability.

Interestingly the number of networks offline was less than one sixth that affected by the blackout, suggesting that organizations have taken steps to ensure network availability even in cases of significant outages. Of course such availability is a major aspect of the architecture of the Internet, but it requires careful planning and execution to make sure that it will work correctly. Let Proton Associates help you ensure your network remains online and ready to go. Contact us for more information.

Are virtual flaws more critical than regular flaws?

noreply@blogger.com (Seth) — Fri, 12 Sep 2008 21:03:00 +0000

That is a question being discussed these days on the heels of VMware’s announcement of 16 vulnerabilities last week across their entire product line. Any flaw in software is bad, but are bugs in the underlying virtualization platforms really more critical than those of the guest operating system?

Compromising the underlying system is highly unlikely to give you access to perform tasks in the guest sessions, but at the same time it would allow you to see network and storage activity and potentially bring down way more systems than crashing just a single non-virtual server would do. In other words, all bugs are bad but these don’t really seem to me to be more critical than a Windows patch.

Microsoft, Red Hat fight security holes

noreply@blogger.com (Seth) — Tue, 26 Aug 2008 02:16:00 +0000

It has been a rough week for security. Microsoft has finally decided to address a potential hole that folks developing in their .NET platform could expose and Red Hat was the victim of a pretty impressive exploit of their Linux server software. No one is safe, it would seem.

First up, the Microsoft thing. SQL Injection attacks have been in the news off and on for a while now. The basic premise is that a hacker can take advantage of shoddy programming to execute their own SQL query against a server instead of letting the server code drive the database access. The end result is usually bad, especially when tied to a buffer overflow in SQL, of which there have been a few. Microsoft has released a new version of their URLScan tool for web servers to address this vulnerability. Rather than just validating the syntax of the URL or the length of the SQL command, the latest version actually can parse the SQL command and decide whether to run it or not based on parameters you provide as an admin. URLScan has com a long way from the original version, but it is still every bit as valuable for a web server administrator, especially now that the .NET tools expose so much more of the server to attack.

The Red Hat attack comes as a bit of a surprise to the open source community since they generally have seen themselves as above attack. Even more troublesome is that it seems this exploit made it into a public release of the OpenSSH build for Red Hat’s Fedora Linux, despite the peer review process that accompanies open source builds and various other protections that supposedly come with the open source movement, someone managed to get exploits into the SSH code of the Linux distribution, and it allowed hackers to compromise the Red Hat site, among other things. There were also reports of inexplicable server reboots in the wild. Red Hat has revoked the security codes for the affected code and has issued a critical patch for their OS to address the issue.

Just remember, everything is a target these days. Make sure you’re protected. Let Proton Associates know if you need help assessing or securing your systems. We’re here to help.

New Treo out there

noreply@blogger.com (Seth) — Fri, 22 Aug 2008 12:43:00 +0000

For folks who are torn between getting a new iPhone or Blackberry, here’s some good news: There is a new Treo model out that might meet your needs. The Treo Pro was announced earlier this week and is expected to be on sale in Europe at some point in September. Palm couldn’t get a US-based carrier to carry the phone, so if you want one in the US you’ll need to buy an unlocked model – at $549 – and then pop it on to either ATT or T-Mobile, as it appears to be GSM-only. The device will run Windows Mobile 6.1 and was GPS and WiFi, among all the other usual bits for a mobile device.

And if you are still not convinced the Blackberry Bold is being released in Canada this week, so that’s an option, too.

Microsoft fixes virtualization licensing rule

noreply@blogger.com (Seth) — Tue, 19 Aug 2008 20:42:00 +0000

Understanding Microsoft licensing has never been easy. The addition of virtualization into an environment only complicates things, as licenses have to be evaluated against both the physical and the virtual server. One huge problem with Microsoft's licensing in a virtual environment was a limitation that only allowed a license to be moved to different physical hardware once in any 90-day period. So if you chose to use VMotion or any other technology to move a virtual host between two different physical servers on a regular basis you'd be in violation of the licensing agreement or you'd have to buy a lot of extra licenses for the applications that you're running.

Today Microsoft changed that policy. Now there is no more 90-day minimum rule. This is a huge benefit for folks running a virtual infrastructure. There are still some limits to the change, most notably that it requires a volume license, not a retail copy of the application and that it only counts for application servers, not the base OS. Lots more details can be found in this blog post. But if you're running Exchange or SQL in a virtual environment this is good news for you.

Need help figuring out your licensing requirements? We can do that for you.

Is the iPhone really going corporate?

noreply@blogger.com (Seth) — Sat, 16 Aug 2008 23:56:00 +0000

It might be, if the CIO for HSBC’s Australia & New Zealand region is to be believed. The financial giant is apparently considering the iPhone as a replacement to their current RIM/Blackberry solution for corporate connectivity on a global scale.

We are actually reviewing iPhones from a HSBC Group perspective ... and when I say that, I mean globally.

Then again, the same person admitted that the move would be a low-priority decision to make, which should come as no surprise considering the cost for the move. Plus it doesn’t take into account the opinions of users, who generally seem to prefer a RIM device for corporate connectivity, though the iPhone certainly is sexier. And the iPhone still lacks some basic requirements for corporate deployment, like a centralized management console, corporate policy support and other features, which makes it all that much more surprising that someone in a CIO role would be suggesting that it is an option.

It makes for a great news story, but I don’t think that it really has legs at this point. I'm guessing they're pushing for a better contract rate with RIM.

Dell unveils new laptop models

noreply@blogger.com (Seth) — Thu, 14 Aug 2008 13:14:00 +0000

Earlier this week Dell announced the details on the latest models of their Latitude line of laptops. There are seven new models coming out, with one that will cover just about every aspect of the market, from the thin and light to a concept model 17” display, 16G RAM beast. They are talking about up to 19 hour battery life in some configurations, though that doesn’t appear to be in a full Windows OS boot mode, multiple forms of connectivity, from WiMAX to WWAN and multiple colors, because everyone knows that mauve has the most RAM.

These announcements are all well and good, but they aren’t the announcement that the market was really hoping for – the new NetBook option. NetBooks are ultra-light and ultra-portable, with very long battery life even running a full Windows (XP) session, but at the expense of processing power. These are all the rage and folks are justifiably disappointed in their absence from this announcement.

Also, Dell has added a digit to all the model numbers and moved from D to E as the leading character, so these laptops really are the next big thing. Enjoy.

Trouble for the iPhone 3G

noreply@blogger.com (Seth) — Thu, 14 Aug 2008 02:14:00 +0000

More and more news outlets are starting to pick up on the stories of poor network performance from the new Apple 3G iPhones, and it appears that the problem might be in the hardware chipset. Gizmodo had a post about it yesterday and now Ny Teknik, a Swedish technical magazine is claiming to have copies of performance reports from the devices showing that the handsets are not performing up to the 3G specifications. One analyst actually is drawing parallels to the initial Motorola Razr release, which had so many problems that the devices were recalled. That would be very bad for Apple.

Both Apple and AT&T are mum on the issue at this point.

Oh, and Apple has admitted that they put some phone-home software in the device. The phone calls home to make sure that the applications running on the device are safe. Would’ve been nice for them to mention that at some point.

But it works with Exchange now, so everything else is water under the bridge, right??

Time runs out on VMware patch

noreply@blogger.com (Seth) — Wed, 13 Aug 2008 20:50:00 +0000

VMware customers running the latest and greatest patch from the virtualization vendor suffered a blow on Tuesday when development code mistakenly left in the most recent patches expired all licenses on VMware ESX servers. The issue affects VMware ESX and ESXi version 3.5 Update 2.

The license expiration code was a remnant from the development version of the patch and was being used to ensure that pre-release patch code didn’t remain in production following testing. The time bomb was never removed from the version that was published to the public, resulting in these issues.

The only available workaround is to set the clock back on the ESX server, which can be a very big problem if you’re running systems according to best practices, which suggest synchronizing the client clock to the ESX host system. This will effectively set back the clock on every server in an environment. That is really, really bad, and very not recommended. There's also a patch that is supposed to address it that got rushed into production.

Google suffers email outage

noreply@blogger.com (Seth) — Tue, 12 Aug 2008 02:56:00 +0000

In what will certainly be a blow to the reputation of their hosted email services business, Google’s Gmail service suffered a two hour outage Monday afternoon US time. The outage knocked many users (including me!) out of their accounts and prevented access via the web interface.

The outage was caused by the contacts component of the system, according to a post on the Gmail website:

From about 2 p.m. to 4 p.m. Pacific Time today, many Gmail users were unable to access their email. The issue is now resolved. We're very sorry for the interruption in service. The issue was caused by a temporary outage in the contacts system used by Gmail which prevented Gmail from loading properly. All mail is safe, though there may be minor delays with delivery.

Since Google sells this service to many organizations with a three nines uptime guarantee (99.9%), they’re now actually looking at having to pay out on their SLA guarantee. The terms of the guarantee measure the outages on a monthly basis and define the credits in those terms. For the month of August, assuming no other outages, the systems will have been up for 742 of 744 hours, or 99.73%. That’s good for three free days for all customers who claim the credit, as is required by the terms. So if you’re a Google Apps Premier Edition customer, queue up that credit request now.

Oh, and the issue affected internal Google communications, as they run on their own platform. Whoopsie.

Intel announces the next family of processors

noreply@blogger.com (Seth) — Mon, 11 Aug 2008 12:52:00 +0000

This is clearly more of a marketing announcement than at technical announcement, as there is no real technical information in the release. But at least we now know what the next generation of processors are going to be called: the “Intel Core i7 processor.” Just rolls off the tongue, doesn’t it?? The i7 processors will be sold for desktop computers, while presumable the Centrino line will continue to be used for laptops and he Xeon line for servers.

Microsoft to provide more information on patches

noreply@blogger.com (Seth) — Fri, 08 Aug 2008 13:21:00 +0000

Or, more accurately, they will be providing more information about the holes that the patches are meant to fix. In addition to the current scale of Critical or Important the patches will also have an “Exploitability Index” number assigned to the patch. This will be released in advance of the patch and is designed to help administrators determine just how critical a patch really is to apply based on the chances that an exploit will appear and/or case damage to systems.

This move seems to be a very useful step for Microsoft customers, as it focuses not just on the fact that there is a hole, but on the likelihood of that hole causing a problem. Of course, there is also the risk that it will give people cover to hide behind when they choose to not patch their systems regularly, but those are just excuses, not real reasons to not do the work.

The new data will be included starting with the October Patch Tuesday cycle.