SNIA: Avoiding tail latency by failing IO operations on purpose

SNIA logoConsistency and predictability matter. You expect Google to answer your search query within a second. If it takes two seconds, that is slow but ok. Much longer and you will probably hit refresh because ‘it’s broken and maybe that will fix it’.

There are many examples that could substitute the scenario above. Starting a Netflix movie, refreshing your Facebook timeline, or powering on an Azure VM. Or in your business: retrieving an MRI scan or patient data, compiling a 3D model, or listing all POs from last month.

Ensuring your service can meet this demand of predictability and consistency requires a multifaceted approach, both in hardware and procedures. You can have a modern hypervisor environment with fast hardware, but if you allow a substantially lower spec system in the cluster, performance will not be consistent. What happens when a virtual machine moves to the lower spec system and suddenly takes longer to finish a query?

Similarly, in storage, tiering across different disk types helps improve TCO. However, what happens when data trickles down to the slowest tier? Achieving that lower TCO comes with the tradeoff of less latency predictability.

These challenges are not new. If they impact user experience too much, you can usually work around them. For example, ensure your data is moved to a faster tier in time. If you have a lot of budget, maybe forgo the slowest & cheapest NL-SAS tier and stick to SAS & SSD. But what if the source of the latency inconsistency is something internal to a component, like a drive?

Continue reading “SNIA: Avoiding tail latency by failing IO operations on purpose”

Storage Field Day 12: storage drop bears reunited!

SFD LogoI’m excited to announce I’ll be attending Storage Field Day 12! During the event we’ll talk storage technology for three days, starting on March 8th. There’s an impressive line-up of companies and delegates gathering in Silicon Valley and of course we’ll live stream the presentations for the folks back home, who can pitch in over Twitter. Did I mention the line-up of companies already? Oh boy!

Continue reading “Storage Field Day 12: storage drop bears reunited!”

Building a hybrid cloud with Avere

Avere logoThere’s no denying that off-premises cloud services are growing. Just look at the year-to-year growth of big public cloud providers. There’s big potential if you focus on two aspects of cloud. The first is speeding up access to data that is potentially not located in the same city or even geographic area. The second is supporting new protocols and storage methodologies that are suited for cloud native applications. One player in this area of IT is Avere, which aims to connect on-premises storage and compute to their siblings in the cloud.

Continue reading “Building a hybrid cloud with Avere”

DellEMCWorld 2016 – I’m impressed!

My DellEMCWorld ride from Los Angeles to AustinWhen the news about the Dell and EMC merger became public last year, I was somewhat skeptical. I’ve had some really sketchy experiences with Dell servers and storage products, so it didn’t feel like a step forward. At the same time there was the organizational and support aspect. Mergers usually result in confusion for both sales processes and us people in the field having to glue all the products together. Not something I was looking forward to.

Lo and behold: I got an invite from the EMC Elect program to attend DellEMCWorld in Austin, Texas! This was my chance to fly over there and experience the merger announcements firsthand, plus ask questions. So I did! And I have to say: I was impressed.

Continue reading “DellEMCWorld 2016 – I’m impressed!”

Public cloud IaaS: is it really that cheap?

Iceland Public CloudA long, long time ago when public cloud IaaS (Infrastructure as a Service) was still relatively new I was doing some contract work for a big international company. One of the tasks for the department was an IaaS proof of concept: does offloading servers to the public cloud result in cost savings? Long story short: the PoC was halted after several months because the AWS IaaS offering was prohibitively expensive and inflexible. We had enough systems and data to keep a team of qualified engineers busy and to get good purchasing discounts. An additional problem was the very rigid service catalog: there weren’t that many flavors of machines available back then and custom machines were either not possible or even more expensive.

Fast forward to 2016 and I’m looking into public/private/on-premises IaaS again for a different company. Prices have dropped, but there are still some things to keep in mind when considering a move to the public cloud IaaS models.

Continue reading “Public cloud IaaS: is it really that cheap?”

Making life a whole lot easier with Tintri VM-aware storage

Tintri IO BlenderAccording to Tintri, the rise of server virtualization broke the traditional storage system. Initially we had relatively simple environments where one server talks to a number of LUNs on a storage system. Sometimes we’d have a small cluster of servers accessing those volumes. Still relatively simple.

Fast forward to now: large clusters of hypervisor hosts are the norm, collectively accessing an even larger number of volumes. Each hypervisor in turn hosts a large number or virtual machines. In case of performance problems, how are you ever going to figure out the root cause and which other systems are affected?

Continue reading “Making life a whole lot easier with Tintri VM-aware storage”

Squashing assumptions with Data Science

Nimble Storage InfoSightEarlier this year Nimble Storage announced their all-flash array called the Predictive Flash Platform; you can read my thoughts on the launch over here. InfoSight is one of the core components of that announcement, which is why we had the opportunity for a fireside chat with the Nimble Storage data science team. We discussed the workings of InfoSight & VMVision and how this relates to actual benefits for an owner of a Nimble Storage array. This post will also touch on some of the key points discussed during the later Storage Field Day 10.

Continue reading “Squashing assumptions with Data Science”

SvSAN 6: now with memory and SSD caching

StorMagic SvSAN schematicA couple of weeks ago StorMagic announced their newest SvSAN 6 release. The basics are still the same: SvSAN takes the internal disks from two hypervisor servers (HyperV or VMware) and turns them into highly available shared storage. Yes, that’s a two server minimum, not three; so this should be a little bit cheaper compared to VMware VSAN and the likes. What’s new in version 6 is the addition of an Advanced edition with SSD and memory-based caching and tiering.

Continue reading “SvSAN 6: now with memory and SSD caching”

FlashBlade: custom hardware still makes sense

Question: What do you get when Pure Storage gets to build a system that can start small, grow big, handle file requests quickly and is simple to manage?

PureStorage FlashBlade

FlashBlade: Pure’s newest addition to its hardware portfolio. The Pure Storage FlashBlade is not just another NAS filer. It’s an all-flash, scale-out storage for file (NFSv3 for now) and object (soon), delivering some pretty good performance as you can see in the sheet above. And the chassis just looks sexy…

Continue reading “FlashBlade: custom hardware still makes sense”

Cloudian HyperStore: manage more PBs with less FTE

Cloudian LogoIf you want to build a private S3 object store, Cloudian HyperStore might be the product for you. Using commodity servers to form a scale-out architecture, you can build your own, fully S3 compliant object storage that’s located in your own datacenter. If you don’t want to supply your own servers, you can opt for the Lenovo Storage DX8200C appliance, powered by Cloudian!

Continue reading “Cloudian HyperStore: manage more PBs with less FTE”